{"year":"2021","title":"'I'm just feeling like it'. On the relationship between the use of the progressive and sentiment polarity in Italian","authors":["L Viola"],"snippet":"… art transformer-based machine learning model for emotion and sentiment classification in Italian which employs the Italian BERT model UmBERTo trained on Commoncrawl ITA (Parisi, Francia, and Magnani [2020] 2021). For …","url":["https://www.uib.no/sites/w3.uib.no/files/attachments/viola.pdf"]}
{"year":"2021","title":"4. Unlocking value from AI in financial services: strategic and organizational tradeoffs vs. media narratives","authors":["G Lanzolla, S Santoni, C Tucci - Artificial Intelligence for Sustainable Value Creation, 2021"],"snippet":"Page 87. 4. Unlocking value from AI in financial services: strategic and organizational tradeoffs vs. media narratives Gianvito Lanzolla, Simone Santoni and Christopher Tucci 1. INTRODUCTION In 1955, McCarthy wrote that …","url":["https://books.google.de/books?hl=en&lr=lang_en&id=_9BCEAAAQBAJ&oi=fnd&pg=PA70&dq=commoncrawl&ots=Z-4LjY9D6U&sig=BHpJ4i9Wq18ZWZDIWoGm5BnHqSY"]}
{"year":"2021","title":"6 Data Collection and Representation for Similar Languages, Varieties and Dialects","authors":["T Samardžic, N Ljubešic - Similar Languages, Varieties, and Dialects: A …, 2021"],"snippet":"… Page 146. Data Collection and Representation for Similar Languages 127 Another project that should be mentioned in this brief overview is the CommonCrawl, a project performing crawls over the whole internet for textual data since 2013 with regular data updates …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=hhA5EAAAQBAJ&oi=fnd&pg=PA121&dq=commoncrawl&ots=2XimIiF4W6&sig=XlIzLoiwAxuodhBmJeC_iS9BSeg"]}
{"year":"2021","title":"\" Short is the Road that Leads from Fear to Hate\": Fear Speech in Indian WhatsApp Groups","authors":["P Saha, B Mathew, K Garimella, A Mukherjee - arXiv preprint arXiv:2102.03870, 2021"],"snippet":"Page 1. “Short is the Road that Leads from Fear to Hate”: Fear Speech in Indian WhatsApp Groups Punyajoy Saha punyajoys@iitkgp.ac.in Indian Institute of Technology Kharagpur, West Bengal, India Binny Mathew binnymathew …","url":["https://arxiv.org/pdf/2102.03870"]}
{"year":"2021","title":"A common framework for quantifying the learnability of nouns and verbs","authors":["Y Zhou, D Yurovsky - Proceedings of the Annual Meeting of the Cognitive …, 2021"],"snippet":"… We used pre-trained 300-dimensional semantic vectors derived from the the Common Crawl corpus composed of 840 billion tokens and 2.2 million words. For our analysis, we considered only the words that corresponded to the relevant 434 images. Procedure …","url":["https://escholarship.org/content/qt8dn6k82j/qt8dn6k82j.pdf"]}
{"year":"2021","title":"A Comparative Study on Word Embeddings in Deep Learning for Text Classification","authors":["C Wang, P Nulty, D Lillis"],"snippet":"… 3https://nlp.stanford.edu/projects/glove/ 4https://commoncrawl.org/ 5https://fasttext.cc/ 6https://allennlp.org/elmo 7We additionally experimented from the fourth-to-last (-4) layer to the last layer … 300s refers to the GloVe …","url":["https://lill.is/pubs/Wang2020a.pdf"]}
{"year":"2021","title":"A Comparison Framework for Product Matching Algorithms","authors":["J Foxcroft - 2021"],"snippet":"Page 1. A Comparison Framework for Product Matching Algorithms by Jeremy Foxcroft A Thesis presented to The University of Guelph In partial fulfilment of requirements for the degree of Master of Science in Computer Science Guelph, Ontario, Canada …","url":["https://atrium.lib.uoguelph.ca/xmlui/bitstream/handle/10214/26375/Foxcroft_Jeremy_202109_Msc.pdf?sequence=3"]}
{"year":"2021","title":"A Comparison of Approaches to Document-level Machine Translation","authors":["Z Ma, S Edunov, M Auli - arXiv preprint arXiv:2101.11040, 2021"],"snippet":"… WMT17 English-German (en-de). For this benchmark, we follow the setup of Müller et al. (2018) whose training data includes the Europarl, Common Crawl, News Commentary and Rapid corpora, totaling nearly 6M sentence pairs …","url":["https://arxiv.org/pdf/2101.11040"]}
{"year":"2021","title":"A Comprehensive Assessment of Dialog Evaluation Metrics","authors":["YT Yeh, M Eskenazi, S Mehri - arXiv preprint arXiv:2106.03706, 2021"],"snippet":"… RoBERTa, which is employed in USR (Mehri and Eskenazi, 2020b), improves the training techniques in BERT and trains the model on a much larger corpus which includes the CommonCrawl News dataset (Mackenzie et al., 2020) and text ex- tracted from Reddit …","url":["https://arxiv.org/pdf/2106.03706"]}
{"year":"2021","title":"A Comprehensive Survey and Experimental Comparison of Graph-Based Approximate Nearest Neighbor Search","authors":["M Wang, X Xu, Q Yue, Y Wang - arXiv preprint arXiv:2101.12631, 2021"],"snippet":"Page 1. A Comprehensive Survey and Experimental Comparison of Graph-Based Approximate Nearest Neighbor Search Mengzhao Wang1, Xiaoliang Xu1, Qiang Yue1, Yuxiang Wang1,∗ 1Hangzhou Dianzi University …","url":["https://arxiv.org/pdf/2101.12631"]}
{"year":"2021","title":"A Comprehensive Survey of Grammatical Error Correction","authors":["Y Wang, Y Wang, K Dang, J Liu, Z Liu - ACM Transactions on Intelligent Systems and …, 2021"],"snippet":"Grammatical error correction (GEC) is an important application aspect of natural language processing techniques, and GEC system is a kind of very important intelligent system that has long been explored both in academic and industrial …","url":["https://dl.acm.org/doi/abs/10.1145/3474840"]}
{"year":"2021","title":"A Computational Framework for Slang Generation","authors":["Z Sun, R Zemel, Y Xu - arXiv preprint arXiv:2102.01826, 2021"],"snippet":"… To compare with and compute the baseline em- bedding methods M for definition sentences, we used 300-dimensional fastText embeddings (Bo- janowski et al., 2017) pre-trained with subword information on 600 billion …","url":["https://arxiv.org/pdf/2102.01826"]}
{"year":"2021","title":"A coral-reef approach to extract information from HTML tables","authors":["P Jiménez Aguirre, JC Roldán Salvador… - Applied Soft Computing …, 2022","P Jiménez, JC Roldán, R Corchuelo - Applied Soft Computing, 2021"],"snippet":"… Unfortunately, a recent analysis of the 32.04 million domains in the November 2019 Common Crawl has revealed that only 11.92 million domains provide such semantic hints [10], which argues for a method to deal with the remaining 20.12 …","url":["https://idus.us.es/bitstream/handle/11441/131990/1/1-s2.0-S1568494621009029-main.pdf?sequence=1","https://www.sciencedirect.com/science/article/pii/S1568494621009029"]}
{"year":"2021","title":"A COVID-19 news coverage mood map of Europe","authors":["F Robertson, J Lagus, K Kajava - Proceedings of the EACL Hackashop on News …, 2021"],"snippet":"… Newscrawl is a web crawl provided by the Common Crawl organisation which is updated more frequently and contains only data from news websites2. In order to keep the size of the corpus manageable and the extraction …","url":["https://www.aclweb.org/anthology/2021.hackashop-1.15.pdf"]}
{"year":"2021","title":"A data quality approach to the identification of discrimination risk in automated decision making systems","authors":["A Vetrò, M Torchiano, M Mecati - Government Information Quarterly, 2021"],"snippet":"… Similarly, a scientific experiment on the search engine Common Crawl (De-Arteaga et al., 2019) revealed an unequal treatment due to gender imbalance in the input data (almost 400,000 biographies): authors compared …","url":["https://www.sciencedirect.com/science/article/pii/S0740624X21000551"]}
{"year":"2021","title":"A data-centric review of deep transfer learning with applications to text data","authors":["S Bashath, N Perera, S Tripathi, K Manjang, M Dehmer… - Information Sciences, 2021"],"snippet":"Abstract In recent years, many applications are using various forms of deep learning models. Such methods are usually based on traditional learning paradigms requiring the consistency of properties among the feature spaces of the training and …","url":["https://www.sciencedirect.com/science/article/pii/S002002552101183X"]}
{"year":"2021","title":"A deep learning-based bilingual Hindi and Punjabi named entity recognition system using enhanced word embeddings","authors":["A Goyal, V Gupta, M Kumar - Knowledge-Based Systems, 2021"],"snippet":"… Initially, we collect Facebook’s pre-trained FastText embeddings which are trained on Wikipedia and common crawl data with 300 dimensions for our Hindi and Punjabi datasets. But after experiments, we find many of the words in our dataset are …","url":["https://www.sciencedirect.com/science/article/pii/S0950705121008637"]}
{"year":"2021","title":"A Framework for Generating Extractive Summary from Multiple Malayalam Documents","authors":["K Manju, S David Peter, SM Idicula - Information, 2021"],"snippet":"… Semantically similar words are mapped to nearby points in the vector space. In this work the vectorization of the terms in the document are performed using the pretrained word embedding model FastText for Malayalam, trained on Common Crawl and Wikipedia …","url":["https://www.mdpi.com/2078-2489/12/1/41/pdf"]}
{"year":"2021","title":"A Framework for Quality Assessment of Semantic Annotations of Tabular Data","authors":["R Avogadro, M Cremaschi, E Jiménez-Ruiz, A Rula - International Semantic Web …, 2021"],"snippet":"… 1 Introduction. Much information is conveyed within tables. A prominent example is the large set of relational databases or tabular data present on the Web. To size the spread of tabular data, 2.5M tables have been …","url":["https://link.springer.com/chapter/10.1007/978-3-030-88361-4_31"]}
{"year":"2021","title":"A Fusion Approach for Paper Submission Recommendation System","authors":["ST Huynh, N Dang, PT Huynh, DH Nguyen, BT Nguyen - International Conference on …, 2021"],"snippet":"… Finally, we use crawl-300d-2M 3 as the pre-train embedding matrix, which has 600 billion tokens and 2 million word vectors trained on Common Crawl. It can make using crawl-300d-2M more efficiently in vectorization. As depicted in Fig …","url":["https://link.springer.com/chapter/10.1007/978-3-030-79463-7_7"]}
{"year":"2021","title":"A General Language Assistant as a Laboratory for Alignment","authors":["A Askell, Y Bai, A Chen, D Drain, D Ganguli… - arXiv preprint arXiv …, 2021"],"snippet":"… For language model pre-training, these models are trained for 400B tokens on a distribution consisting mostly of filtered common crawl … The natural language dataset was composed of 55% heavily filtered common crawl data (220B tokens), 32 …","url":["https://arxiv.org/pdf/2112.00861"]}
{"year":"2021","title":"A Heuristic-driven Ensemble Framework for COVID-19 Fake News Detection","authors":["SD Das, A Basak, S Dutta - arXiv preprint arXiv:2101.03545, 2021"],"snippet":"… of model-specific special tokens. Each model also has its corresponding vocabulary associated with its tokenizer, trained on a large corpus data like GLUE, wikitext-103, CommonCrawl data etc. During training, each model …","url":["https://arxiv.org/pdf/2101.03545"]}
{"year":"2021","title":"A Heuristic-driven Uncertainty based Ensemble Framework for Fake News Detection in Tweets and News Articles","authors":["SD Das, A Basak, S Dutta - arXiv preprint arXiv:2104.01791, 2021"],"snippet":"… Each model also has its corresponding vocabulary associated with its tokenizer, trained on a large corpus data like GLUE, wikitext-103, CommonCrawl data etc. During training, each model applies the tokenization …","url":["https://arxiv.org/pdf/2104.01791"]}
{"year":"2021","title":"A Human Being Wrote This Law","authors":["AB Cyphert"],"snippet":"… GPT-3 had an impressively large data training set: it was trained on the Common Crawl dataset, a nearly trillion-word dataset,22 which includes everything from traditional news sites like the New York Times to sites like Reddit.The Common …","url":["https://lawreview.law.ucdavis.edu/issues/55/1/articles/files/55-1_Cyphert.pdf"]}
{"year":"2021","title":"A Literature Survey of Recent Advances in Chatbots","authors":["G Caldarini, S Jaf, K McGarry - 2021"],"snippet":"… This led to the development of pretrained systems such as BERT (Bidirectional Encoder Representations from transformers) [46] and GPT (Generative Pre-trained Transformer), which were trained with huge language datasets, such as Wikipedia …","url":["https://www.preprints.org/manuscript/202112.0265/download/final_file"]}
{"year":"2021","title":"A Mechanism for Producing Aligned Latent Spaces with Autoencoders","authors":["S Jain, A Radhakrishnan, C Uhler - arXiv preprint arXiv:2106.15456, 2021"],"snippet":"… 6.1 Alignment of GloVe Embeddings In this section, we apply our theory to align semantic/syntactic directions in GloVe word embeddings [21]. We use 300 dimensional GloVe vectors that were trained on Common Crawl with 840 billion tokens …","url":["https://arxiv.org/pdf/2106.15456"]}
{"year":"2021","title":"A Multi-Platform Analysis of Political News Discussion and Sharing on Web Communities","authors":["Y Wang, S Zannettou, J Blackburn, B Bradlyn… - arXiv preprint arXiv …, 2021"],"snippet":"… supported types of entities). The model re- lies on Convolutional Neural Networks (CNNs), trained on the OntoNotes dataset [90], as well as Glove vectors [62] trained on the Common Crawl dataset [17]. 2.3 News Stories Identification …","url":["https://arxiv.org/pdf/2103.03631"]}
{"year":"2021","title":"A Multi-Task Learning Model for Multidimensional Relevance Assessment","authors":["DGP Putri, M Viviani, G Pasi - International Conference of the Cross-Language …, 2021"],"snippet":"… 6 In particular, we focused on the ad-hoc retrieval subtask. The data consist of Web pages crawled by means of CommonCrawl, 7 related to the health-related domain. The data collections consider 50 topics/queries and associated documents …","url":["https://link.springer.com/chapter/10.1007/978-3-030-85251-1_9"]}
{"year":"2021","title":"A Multifactorial Approach to Crosslinguistic Constituent Orderings","authors":["Z Liu"],"snippet":"… The data for training these LMs was taken from the raw data of the CoNLL 2017 Shared Task on multilingual parsing (Ginter et al. 2017), which contains texts from Common Crawl and Wikipedia. The architecture of the LM was the same for every language …","url":["https://www.researchgate.net/profile/Zoey-Liu/publication/354204297_A_Multifactorial_Approach_to_Crosslinguistic_Constituent_Orderings/links/612c0095c69a4e487967c628/A-Multifactorial-Approach-to-Crosslinguistic-Constituent-Orderings.pdf"]}
{"year":"2021","title":"A Multitask Framework to Detect Depression, Sentiment and Multi-label Emotion from Suicide Notes","authors":["S Ghosh, A Ekbal, P Bhattacharyya - Cognitive Computation, 2021"],"snippet":"The significant rise in suicides is a major cause of concern in public health domain. Depression plays a major role in increasing suicide ideation among th.","url":["https://link.springer.com/article/10.1007/s12559-021-09828-7"]}
{"year":"2021","title":"A Novel Corpus of Discourse Structure in Humans and Computers","authors":["B Hemmatian, S Feucht, R Avram, A Wey, M Garg… - arXiv preprint arXiv …, 2021"],"snippet":"We present a novel corpus of 445 humanand computer-generated documents, comprising about 27,000 clauses, annotated for semantic clause types and coherence relations that allow for nuanced comparison of artificial and natural …","url":["https://arxiv.org/pdf/2111.05940"]}
{"year":"2021","title":"A novel fusion-based deep learning model for sentiment analysis of COVID-19 tweets","authors":["ME Basiri, S Nemati, M Abdar, S Asadi, UR Acharrya - Knowledge-Based Systems, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0950705121005049"]}
{"year":"2021","title":"A NOVEL TRILINGUAL DATASET FOR CRISIS NEWS CATEGORIZATION ACROSS LANGUAGES","authors":["K Kajava - 2021"],"snippet":"… Use Common Crawl instead • “the goal of democratizing access to web information by producing and maintaining an open repository of web crawl data that is universally accessible and analyzable” (https://commoncrawl.org/about/, accessed June 1 2021) …","url":["https://blogs.helsinki.fi/language-technology/files/2021/06/LT-Seminar-2021-06-03-Kaisla-Kajava.pdf"]}
{"year":"2021","title":"A Primer on Pretrained Multilingual Language Models","authors":["S Doddapaneni, G Ramesh, A Kunchukuttan, P Kumar… - arXiv preprint arXiv …, 2021"],"snippet":"… and they differ in the architecture (eg, number of layers, parameters, etc), objective functions used for training (eg, monolingual masked language modeling objective, translation language modeling objective, etc), data used …","url":["https://arxiv.org/pdf/2107.00676"]}
{"year":"2021","title":"A Probing Task on Linguistic Properties of Korean Sentence Embedding","authors":["A Ahn, BI Ko, D Lee, G Han, M Shin, J Nam - Annual Conference on Human and …, 2021"],"snippet":"Abstract 본 연구는 한국어 문장 임베딩 (embedding) 에 담겨진 언어적 속성을 평가 하기 위한 프로빙 태스크 (Probing Task) 를 소개한다. 프로빙 태스크는 임베딩 으로부터 문장의 표층적, 통사적, 의미적 속성을 구분하는 문제로 영어, 폴란드어 …","url":["https://www.koreascience.or.kr/article/CFKO202130060614813.pdf"]}
{"year":"2021","title":"A Residual Network Architecture for Hindi NER using Fasttext and BERT embedding layers","authors":["R Shelke, S Vanjale"],"snippet":"… It provides word embedding for Hindi (and 157 other languages) and is based on the CBOW (Continuous Bag-of-Words) model. The CBOW model learns by predicting the current word based on its context, and it was trained on Common Crawl and Wikipedia …","url":["https://www.novyimir.net/gallery/nmrj%202867f.pdf"]}
{"year":"2021","title":"A Review of Bangla Natural Language Processing Tasks and the Utility of Transformer Models","authors":["F Alam, A Hasan, T Alam, A Khan, J Tajrin, N Khan… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. A Review of Bangla Natural Language Processing Tasks and the Utility of Transformer Models FIROJ ALAM, Qatar Computing Research Institute, HBKU, Qatar MD. ARID HASAN, Cognitive Insight Limited, Bangladesh …","url":["https://arxiv.org/pdf/2107.03844"]}
{"year":"2021","title":"A Review of Public Datasets in Question Answering Research","authors":["BB Cambazoglu, M Sanderson, F Scholer, B Croft"],"snippet":"… matching the question. The sentences are selected based on their tf-idf similarity to the question. The underlying web page collection contains pages from the July 2018 archive of the Common Crawl web repository. Task. Given a …","url":["http://www.sigir.org/wp-content/uploads/2020/12/p07.pdf"]}
{"year":"2021","title":"A Semi-supervised Multi-task Learning Approach to Classify Customer Contact Intents","authors":["L Dong, MC Spencer, A Biagi"],"snippet":"… We note that this ALBERT model is trained as a multiclass classification with only positive cases. 3.2.2 SS MT D/TAPT ALBERT The pretrained language models are mostly trained on well-known corpora, such as Wikipedia, Common Crawl, BookCorpus, Reddit, etc …","url":["https://assets.amazon.science/79/22/d9237534448293405083a73b896d/a-semi-supervised-multi-task-learning-approach-to-classify-customer-contact-intents.pdf"]}
{"year":"2021","title":"A Short Survey of LSTM Models for De-identification of Medical Free Text","authors":["JL Leevy, TM Khoshgoftaar - 2020 IEEE 6th International Conference on …, 2020"],"snippet":"… The training set was obtained from the 2014 i2b2 challenge, while the test set came from the University of Florida (UF) Health Integrated Data Repository 5. Word embeddings were sourced from GoogleNews [54] …","url":["https://ieeexplore.ieee.org/abstract/document/9319017/"]}
{"year":"2021","title":"A Simple Post-Processing Technique for Improving Readability Assessment of Texts using Word Mover's Distance","authors":["JM Imperial, E Ong - arXiv preprint arXiv:2103.07277, 2021"],"snippet":"… technique described in Section 4. For the word embeddings of English, German, and Filipino needed for the technique, we downloaded the resources from the fastText website5. The word embeddings in various …","url":["https://arxiv.org/pdf/2103.07277"]}
{"year":"2021","title":"A Simple Recipe for Multilingual Grammatical Error Correction","authors":["S Rothe, J Mallinson, E Malmi, S Krause, A Severyn - arXiv preprint arXiv:2106.03830, 2021"],"snippet":"… 2.1 mT5 Pre-training mT5 has been pre-trained on mC4 corpus, a subset of Common Crawl, covering 101 languages and composed of about 50 billion documents. For details on mC4, we refer the reader to the original paper (Xue et al., 2020) …","url":["https://arxiv.org/pdf/2106.03830"]}
{"year":"2021","title":"A Spontaneous Stereotype Content Model: Taxonomy, Properties, and Prediction.","authors":["G Nicolas, X Bai, ST Fiske"],"snippet":"… model trained on the Common Crawl (600 billion words obtained from various internet sources) … the Common Crawl (600 billion words), a Glove model trained using around 840 billion words from the Common Crawl (Pennington, Socher, & Manning, 2014; …","url":["https://www.nicolaslab.org/publication/sscm/SSCM.pdf"]}
{"year":"2021","title":"A Study of Analogical Density in Various Corpora at Various Granularity","authors":["R Fam, Y Lepage - Information, 2021"],"snippet":"In this paper, we inspect the theoretical problem of counting the number of analogies between sentences contained in a text. Based on this, we measure the analogical density of the text. We focus on analogy at the sentence level …","url":["https://www.mdpi.com/2078-2489/12/8/314/pdf"]}
{"year":"2021","title":"A Study of Analogical Density in Various Corpora at Various Granularity. Information 2021, 12, 314","authors":["R Fam, Y Lepage - 2021"],"snippet":"… Table 4 shows the statistics of Multi30K corpus. • CommonCrawl (available at: commoncrawl.org accessed on 20 September 2020) is a crawled web archive and dataset … Table 5 shows the statistics on the CommonCrawl corpus …","url":["https://search.proquest.com/openview/208b192bc36d7c71728c73989c304dea/1?pq-origsite=gscholar&cbl=2032384"]}
{"year":"2021","title":"A study on performance improvement considering the balance between corpus in Neural Machine Translation","authors":["C Park, K Park, H Moon, S Eo, H Lim - Journal of the Korea Convergence Society, 2021"],"snippet":"… 1. Concept of Corpus Weight Balance GPT3도 Common Crawl, WebText2, Books1, Books2, Wikipedia 등의 데이터를 합쳐 모델을 훈련하 게 된다. 그러나 말뭉치 간의 특성 및 특징(어투, 문체, 도메인 등)이 다름에도 하나의 데이터로 …","url":["https://www.koreascience.or.kr/article/JAKO202116954598769.pdf"]}
{"year":"2021","title":"A Survey of COVID-19 Misinformation: Datasets, Detection Techniques and Open Issues","authors":["AR Ullah, A Das, A Das, MA Kabir, K Shu - arXiv preprint arXiv:2110.00737, 2021"],"snippet":"Page 1. A Survey of COVID-19 Misinformation: Datasets, Detection Techniques and Open Issues AR Sana Ullaha, Anupam Dasa, Anik Dasb, Muhammad Ashad Kabirc,∗, Kai Shud aDepartment of Computer Science and Engineering …","url":["https://arxiv.org/pdf/2110.00737"]}
{"year":"2021","title":"A Survey of Machine Learning-Based Solutions for Phishing Website Detection","authors":["L Tang, QH Mahmoud - Machine Learning and Knowledge Extraction, 2021"],"snippet":"With the development of the Internet, network security has aroused people's attention. It can be said that a secure network environment is a basis for the rapid and sound development of the Internet. Phishing is an essential class …","url":["https://www.mdpi.com/2504-4990/3/3/34/pdf"]}
{"year":"2021","title":"A Survey of Recent Abstract Summarization Techniques","authors":["D Puspitaningrum - Proceedings of Sixth International Congress on …, 2021"],"snippet":"… For C4, taken from Common Crawl scrape from April 2019 and applied some cleansing filters, it results in a very clean 750GB text dataset of large pre-training datasets, more extensive than other pre-training datasets. 3.2 Pegasus-XSum (Pegasus) …","url":["https://hal.archives-ouvertes.fr/hal-03216381/document"]}
{"year":"2021","title":"A Survey on Bias in Deep NLP","authors":["I Garrido-Muñoz, A Montejo-Ráez, F Martínez-Santiago… - 2021"],"snippet":"… 2016 [26] Gender Word2Vec, GloVe GoogleNews corpus (w2vNEWS), Common Crawl English Analogies/Cosine Similarity Vector Space Manipulation After - 2017 [10] Gender, Ethnicity GloVe, Word2Vec Common …","url":["https://www.preprints.org/manuscript/202103.0049/download/final_file"]}
{"year":"2021","title":"A Survey on Data Augmentation for Text Classification","authors":["M Bayer, MA Kaufhold, C Reuter - arXiv preprint arXiv:2107.03158, 2021"],"snippet":"… CNN+LSTM/GRU HON RSN-1 RSN-2 Word2Vec Hate Speech FastText Wikipedia GoogleNews W2V GloVe Common Crawl GloVe Common Crawl GloVe Common Crawl -22.7 (Macro F1) +1.0 -3.3 +0.3 -0.2 0 [44] 1. Method …","url":["https://arxiv.org/pdf/2107.03158"]}
{"year":"2021","title":"A Survey on Low-Resource Neural Machine Translation","authors":["R Wang, X Tan, R Luo, T Qin, TY Liu - arXiv preprint arXiv:2107.04239, 2021"],"snippet":"Page 1. A Survey on Low-Resource Neural Machine Translation Rui Wang, Xu Tan, Renqian Luo, Tao Qin and Tie-Yan Liu Microsoft Research Asia {ruiwa, xuta, t-reluo, taoqin, tyliu}@microsoft.com Abstract Neural approaches …","url":["https://arxiv.org/pdf/2107.04239"]}
{"year":"2021","title":"A Survey On Neural Word Embeddings","authors":["E Sezerer, S Tekir - arXiv preprint arXiv:2110.01804, 2021"],"snippet":"… ivLBL/vLBL [95] 2013 100-600 Wiki LBL Performance - NCE [47] GloVe [109] 2014 300 Wiki, Gigaword, Commoncrawl LBL+coocurence Matrix Training - - DEPS [69] 2014 300 Wiki CBOW Training Stanford tagger[129] …","url":["https://arxiv.org/pdf/2110.01804"]}
{"year":"2021","title":"A Survey on Statistical Approaches for Abstractive Summarization of Low Resource Language Documents","authors":["P Deshpande, S Jahirabadkar - Smart Trends in Computing and Communications, 2022"],"snippet":"… German Wiki data is used as real data and synthetic data is a common crawl data. Synthetic data is used to increase size of data. Three settings are considered for generation of summaries: (1) Transformer model using real data for training. …","url":["https://link.springer.com/chapter/10.1007/978-981-16-4016-2_69"]}
{"year":"2021","title":"A Synthetic FACS Framework for Expanding Facial Expression Lexicons","authors":["C Butler - 2021"],"snippet":"Page 1. A Synthetic FACS Framework for Expanding Facial Expression Lexicons DISSERTATION Submitted in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY (Computer Science) at the …","url":["https://search.proquest.com/openview/ebf5a87df8275e6fc6fac0b1c0b21b44/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"A system for proactive risk assessment of application changes in cloud operations","authors":["R Batta, L Shwartz, M Nidd, AP Azad, H Kumar - 2021 IEEE 14th International …, 2021"],"snippet":"Abstract Change is one of the biggest contributors to service outages. With more enterprises migrating their applications to cloud and using automated build and deployment the volume and rate of changes has significantly increased. Furthermore …","url":["https://www.computer.org/csdl/proceedings-article/cloud/2021/006000a112/1ymJ4TXNxUA"]}
{"year":"2021","title":"A Systematic Investigation of Commonsense Understanding in Large Language Models","authors":["XL Li, A Kuncoro, CM d'Autume, P Blunsom… - arXiv preprint arXiv …, 2021"],"snippet":"… 2019), we train our models using the cleaned version of Common Crawl corpus (C4), around 800 GB of data. Our largest model, with 32 transformer layers and 7 billion parameters, has a similar number of parameters to the open-sourced GPT-J model (Wang …","url":["https://arxiv.org/pdf/2111.00607"]}
{"year":"2021","title":"A systems-wide understanding of the human olfactory percept chemical space","authors":["J Kowalewski, B Huynh, A Ray - Chemical Senses, 2021"],"snippet":"… 2015; spaCy, 2016), and a convolutional neural network previously trained on GloVe Common Crawl (Pennington, Socher, & Manning, 2014) and OntoNotes 5. The training set is comprised of more than 1 million English …","url":["https://academic.oup.com/chemse/advance-article-abstract/doi/10.1093/chemse/bjab007/6153471"]}
{"year":"2021","title":"A Targeted Attack on Black-Box Neural Machine Translation with Parallel Data Poisoning","authors":["C Xu, J Wang, Y Tang, F Guzmán, BIP Rubinstein… - Proceedings of the Web …, 2021"],"snippet":"… 3https://commoncrawl.org/ 4We assume that these poisoned web pages are archived and to be used for parallel data extraction. This assumption is realistic as we find that the crawling services commonly used for parallel …","url":["https://dl.acm.org/doi/abs/10.1145/3442381.3450034"]}
{"year":"2021","title":"A unified approach to sentence segmentation of punctuated text in many languages","authors":["R Wicks, M Post"],"snippet":"Page 1. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 3995–4007 August 1–6, 2021 …","url":["https://aclanthology.org/2021.acl-long.309.pdf"]}
{"year":"2021","title":"A word embedding-based approach to cross-lingual topic modeling","authors":["CH Chang, SY Hwang - Knowledge and Information Systems, 2021"],"snippet":"The cross-lingual topic analysis aims at extracting latent topics from corpora of different languages. Early approaches rely on high-cost multilingual reso.","url":["https://link.springer.com/article/10.1007/s10115-021-01555-7"]}
{"year":"2021","title":"ABC: Attention with Bounded-memory Control","authors":["H Peng, J Kasai, N Pappas, D Yogatama, Z Wu, L Kong… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Under review as a conference paper at ICLR 2022 ABC: ATTENTION WITH BOUNDED-MEMORY CONTROL Hao Peng♠ Jungo Kasai♠ Nikolaos Pappas♠ Dani Yogatama♣ Zhaofeng Wu♦∗ Lingpeng Kong♦ Roy Schwartz …","url":["https://arxiv.org/pdf/2110.02488"]}
{"year":"2021","title":"Abuse is Contextual, What about NLP? The Role of Context in Abusive Language Annotation and Detection","authors":["S Menini, AP Aprosio, S Tonelli - arXiv preprint arXiv:2103.14916, 2021"],"snippet":"… In particular, all vectors are extracted starting from the pre-trained embeddings obtained from the Common Crawl corpus.5 Since SVM takes in input sentence embeddings, we convert the context and the current tweet …","url":["https://arxiv.org/pdf/2103.14916"]}
{"year":"2021","title":"Accelerated execution via eager-release of dependencies in task-based workflows","authors":["H Elshazly, F Lordan, J Ejarque, RM Badia - The International Journal of High …, 2021"],"snippet":"Task-based programming models offer a flexible way to express the unstructured parallelism patterns of nowadays complex applications. This expressive capability is required to achieve maximum possi...","url":["https://journals.sagepub.com/doi/abs/10.1177/1094342021997558"]}
{"year":"2021","title":"Accelerating Text Communication via Abbreviated Sentence Input","authors":["J Adhikary, J Berger, K Vertanen"],"snippet":"… For our out-of-domain training set, we used one billion words of web text from Common Crawl1. We only … As shown in Table 1, random sentences from Common Crawl averaged 30 words. The cross-entropy 1https://commoncrawl …","url":["https://aclanthology.org/2021.acl-long.514.pdf"]}
{"year":"2021","title":"Accurate Word Representations with Universal Visual Guidance","authors":["Z Zhang, H Yu, H Zhao, R Wang, M Utiyama - arXiv preprint arXiv:2012.15086, 2020"],"snippet":"… WMT'14 EN-DE 4.43M bilingual sentence pairs of the WMT14 dataset were used as training data, including Common Crawl, News Commentary, and Europarl v7. The newstest2013 and newstest2014 datasets were …","url":["https://arxiv.org/pdf/2012.15086"]}
{"year":"2021","title":"Acquiring and Harnessing Verb Knowledge for Multilingual Natural Language Processing","authors":["O Majewska - 2021"],"snippet":"Advances in representation learning have enabled natural language processing models to derive non-negligible linguistic information directly from text corpora in an unsupervised fashion. However, this signal is underused in downstream tasks …","url":["https://www.repository.cam.ac.uk/bitstream/handle/1810/329292/Majewska_PhDThesis_final.pdf?sequence=4"]}
{"year":"2021","title":"Active Learning for Argument Mining: A Practical Approach","authors":["N Solmsdorf, D Trautmann, H Schütze - arXiv preprint arXiv:2109.13611, 2021"],"snippet":"… easily discernible argumentative statements. The corpus contains 1,000 sentences per topic, ie, in total 8,000 instances, which were tapped from a Common Crawl snapshot and in- dexed with Elasticsearch. The time-consuming …","url":["https://arxiv.org/pdf/2109.13611"]}
{"year":"2021","title":"Adapting Neural Machine Translation for Automatic Post-Editing","authors":["A Sharma, P Gupta, A Nelakanti"],"snippet":"… reference as the output. 3.2 Pre-training on domain-specific data FAIR's WMT'19 NMT model was trained on Newscrawl and Commoncrawl datasets while the source of this year's APE data is Wikipedia. To fix the domain mismatch …","url":["https://assets.amazon.science/dc/df/5443c00541a9b6257f6110c5bb86/adapting-neural-machine-translation-for-automatic-post-editing.pdf"]}
{"year":"2021","title":"Adaptive Ranking Relevant Source Files for Bug Reports Using Genetic Algorithm","authors":["H Fujita, H Perez-Meana - 2021"],"snippet":"Abstract. Precisely locating buggy files for a given bug report is a cumbersome and time-consuming task, particularly in a large-scale project with thousands of source files and bug reports. An efficient bug localization module is desirable to improve the …","url":["https://books.google.de/books?hl=en&lr=lang_en&id=GYxJEAAAQBAJ&oi=fnd&pg=PA430&dq=commoncrawl&ots=IRy65Bdasc&sig=WyEQAu3sL155gzwLW6pjIX4mCwQ"]}
{"year":"2021","title":"ADEPT: An Adjective-Dependent Plausibility Task","authors":["A Emami, I Porada, A Olteanu, K Suleman, A Trischler…"],"snippet":"… 4 Dataset To construct ADEPT, we scrape text samples from English Wikipedia and Common Crawl, extracting adjectival modifier-noun pairs that occur with high frequency … We extracted 10 million pairs from English …","url":["https://aclanthology.org/2021.acl-long.553.pdf"]}
{"year":"2021","title":"ADPBC: Arabic Dependency Parsing Based Corpora for Information Extraction","authors":["S Mohamed, M Hussien, HM Mousa - 2021"],"snippet":"… Sch. Econ. Res. Pap. No. WP BRP., 2018. [27] A. Panchenko, E. Ruppert, S. Faralli, SP Ponzetto, and C. Biemann, “Building a web-scale dependency-parsed corpus from common crawl,” Lr. 2018 - 11th Int. Conf. Lang. Resour. Eval., pp. 1816–1823, 2019 …","url":["http://www.mecs-press.org/ijitcs/ijitcs-v13-n1/IJITCS-V13-N1-4.pdf"]}
{"year":"2021","title":"Advances and Trends in Artificial Intelligence. From Theory to Practice: 34th International Conference on Industrial, Engineering and Other Applications of Applied …","authors":["H Fujita"],"snippet":"Page 1. Hamido Fujita Ali Selamat Jerry Chun-Wei Lin Moonis Ali (Eds.) Advances and Trends in Artificial Intelligence From Theory to Practice 34th International Conference on Industrial, Engineering and Other Applications …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=ihg5EAAAQBAJ&oi=fnd&pg=PR5&dq=commoncrawl&ots=fSoeFLOj--&sig=NR8xQHWDWjlGUhvSgV52wEaaT6Y"]}
{"year":"2021","title":"Aggressive and Offensive Language Identification in Hindi, Bangla, and English: A Comparative Study","authors":["R Kumar, B Lahiri, AK Ojha - SN Computer Science, 2021"],"snippet":"In the present paper, we carry out a comparative study between offensive and aggressive language and attempt to understand their inter-relationship. To car.","url":["https://link.springer.com/article/10.1007/s42979-020-00414-6"]}
{"year":"2021","title":"AlephBERT: A Hebrew Large Pre-Trained Language Model to Start-off your Hebrew NLP Application With","authors":["A Seker, E Bandel, D Bareket, I Brusilovsky… - arXiv preprint arXiv …, 2021"],"snippet":"… Oscar: A deduplicated Hebrew portion of the OSCAR corpus, which is “extracted from Common Crawl via language classification, filtering and cleaning” (Ortiz Suárez et al., 2020). • Twitter: Texts of Hebrew tweets collected between 2014-09-28 and 2018-03-07 …","url":["https://arxiv.org/pdf/2104.04052"]}
{"year":"2021","title":"Alignment of Language Agents","authors":["Z Kenton, T Everitt, L Weidinger, I Gabriel, V Mikulik… - arXiv preprint arXiv …, 2021","ZKTEL Weidinger, IGVMG Irving"],"snippet":"… Large scale unlabeled datasets are collected from the web, such as the CommonCrawl dataset (Raffel et al., 2019). Input data and labels are created by chopping a sentence into … Brown et al. (2020) attempt to im- prove …","url":["https://ar5iv.labs.arxiv.org/html/2103.14659","https://arxiv.org/pdf/2103.14659"]}
{"year":"2021","title":"All Labels Are Not Created Equal: Enhancing Semi-supervision via Label Grouping and Co-training","authors":["I Nassar, S Herath, E Abbasnejad, W Buntine, G Haffari - arXiv preprint arXiv …, 2021"],"snippet":"… A detailed description of such relations and examples thereof can be found in the ConceptNet documentation8. On the other hand, GloVe and word2vec are two prominent sets of word embeddings, the former is trained on 840 …","url":["https://arxiv.org/pdf/2104.05248"]}
{"year":"2021","title":"All NLP Tasks Are Generation Tasks: A General Pretraining Framework","authors":["Z Du, Y Qian, X Liu, M Ding, J Qiu, Z Yang, J Tang - arXiv preprint arXiv:2103.10360, 2021"],"snippet":"Page 1. All NLP Tasks Are Generation Tasks: A General Pretraining Framework Zhengxiao Du *12 Yujie Qian * 3 Xiao Liu 1 2 Ming Ding 1 2 Jiezhong Qiu 1 2 Zhilin Yang 4 2 Jie Tang 1 2 Abstract There have been various types …","url":["https://arxiv.org/pdf/2103.10360"]}
{"year":"2021","title":"Allocating Large Vocabulary Capacity for Cross-lingual Language Model Pre-training","authors":["B Zheng, L Dong, S Huang, S Singhal, W Che, T Liu… - arXiv preprint arXiv …, 2021"],"snippet":"… are learned on the reconstructed CommonCrawl corpus (Chi et al., 2021b; Conneau et al., 2020) using SentencePiece (Kudo and Richardson, 2018) with the unigram language model (Kudo, 2018). The unigram distributions …","url":["https://arxiv.org/pdf/2109.07306"]}
{"year":"2021","title":"ALX: Large Scale Matrix Factorization on TPUs","authors":["H Mehta, S Rendle, W Krichene, L Zhang - arXiv preprint arXiv:2112.02194, 2021"],"snippet":"We present ALX, an open-source library for distributed matrix factorization using Alternating Least Squares, written in JAX. Our design allows for efficient use of the TPU architecture and scales well to matrix factorization problems of O(B) rows/columns …","url":["https://arxiv.org/pdf/2112.02194"]}
{"year":"2021","title":"AMMUS: A Survey of Transformer-based Pretrained Models in Natural Language Processing","authors":["KS Kalyan, A Rajasekharan, S Sangeetha - arXiv preprint arXiv:2108.05542, 2021"],"snippet":"… mT6 [91], XLM-E [89] CC-Aligned [108] Parallel corpus of 292 million non-English common crawl document pairs and 100 million English common crawl document pairs. XLM-E [89] Dakshina [109] Parallel corpus containing 10K sentences for 12 In- dian languages …","url":["https://arxiv.org/pdf/2108.05542"]}
{"year":"2021","title":"An Alignment-Based Approach to Semi-Supervised Bilingual Lexicon Induction with Small Parallel Corpora","authors":["K Marchisio, C Xiong, P Koehn"],"snippet":"… learning. 6 Experimental Settings Language Corpus # of words English WaCky, BNC, Wikipedia 2.8 B Italian itWac 1.6 B German SdeWaC 0.9 B Spanish News Crawl 2007-2012 386 M Finnish Common Crawl 2016 2.8 B Table …","url":["https://aclanthology.org/2021.mtsummit-research.24.pdf"]}
{"year":"2021","title":"An analysis of full-size Russian complexly NER labelled corpus of Internet user reviews on the drugs based on deep learning and language neuron nets","authors":["AG Sboeva, SG Sboevac, IA Moloshnikova…"],"snippet":"Page 1. An analysis of full-size Russian complexly NER labelled corpus of Internet user reviews on the drugs based on deep learning and language neuron nets AG Sboeva,b,, SG Sboevac, IA Moloshnikova, AV Gryaznova, RB …","url":["https://sagteam.ru/papers/med-corpus/4.pdf"]}
{"year":"2021","title":"An Effective Deep Learning Approach for Extractive Text Summarization","authors":["MT Luu, TH Le, MT Hoang"],"snippet":"Page 1. An Effective Deep Learning Approach for Extractive Text Summarization Minh-Tuan Luu PhD. Student, School of Information and Communication Technology, Hanoi University of Science and Technology, No.1 Dai Co …","url":["http://www.ijcse.com/docs/INDJCSE21-12-02-141.pdf"]}
{"year":"2021","title":"An embedding method for unseen words considering contextual information and morphological information","authors":["MS Won, YS Choi, S Kim, CW Na, JH Lee - Proceedings of the 36th Annual ACM …, 2021"],"snippet":"… Random embeddings are assigned for OOv words. Glove is implemented by using preǦtrained 300Ǧdimensional Glove embedding, which is trained on Common Crawl with 840B word tokens. Random embeddings are assigned for OOv words …","url":["https://dl.acm.org/doi/abs/10.1145/3412841.3441982"]}
{"year":"2021","title":"An empirical evaluation of text representation schemes to filter the social media stream","authors":["S Modha, P Majumder, T Mandl - Journal of Experimental & Theoretical Artificial …, 2021"],"snippet":"… Glove pre-trained model available with different embed size and trained on the common crawl, Twitter. We have used the Glove pre-trained model with a vocabulary size of 2.2 million and trained on the common crawl. fastText …","url":["https://www.tandfonline.com/doi/full/10.1080/0952813X.2021.1907792"]}
{"year":"2021","title":"An Empirical Exploration in Quality Filtering of Text Data","authors":["L Gao - arXiv preprint arXiv:2109.00698, 2021"],"snippet":"… (2020), with a Paretodistribution thresholded filtering method and a shallow CommonCrawl-WebText classfier … (2020) has been made public, we instead use the same type of fasttext (Joulin et al., 2017) classifier between unfiltered …","url":["https://arxiv.org/pdf/2109.00698"]}
{"year":"2021","title":"An Empirical Study on Task-Oriented Dialogue Translation","authors":["S Liu - ICASSP 2021-2021 IEEE International Conference on …, 2021"],"snippet":"… consistent). We valid them with SENT-BASE model on En⇒De task. data in WMT20 news domain, which consists of CommonCrawl and NewsCommentary. We conduct data selection to select similar amount of sentences …","url":["https://ieeexplore.ieee.org/abstract/document/9413521/"]}
{"year":"2021","title":"An End-to-end Point of Interest (POI) Conflation Framework","authors":["R Low, ZD Tekler, L Cheah - arXiv preprint arXiv:2109.06073, 2021"],"snippet":"… words that did not appear in the training data [63]. For this study, the fastText model was pre-trained on 2 million word vectors with subword information from commoncrawl.org. The second advantage of using the fastText library to …","url":["https://arxiv.org/pdf/2109.06073"]}
{"year":"2021","title":"An evaluation dataset for depression detection in Arabic social media","authors":["S Elimam, M Bougeussa - International Journal of Knowledge Engineering and …, 2021"],"snippet":"Studying depression in Arabic social media has been neglected compared to other languages and the traditional way of dealing with depression (face-to-face medical diagnose) is not enough as the number of people that suffer from depression in …","url":["https://www.inderscienceonline.com/doi/abs/10.1504/IJKEDM.2021.119888"]}
{"year":"2021","title":"An Explainable Multi-Modal Hierarchical Attention Model for Developing Phishing Threat Intelligence","authors":["Y Chai, Y Zhou, W Li, Y Jiang - IEEE Transactions on Dependable and Secure …, 2021"],"snippet":"Phishing website attack, as one of the most persistent forms of cyber threats, evolves and remains a major cyber threat. Various detection methods (eg, lookup systems, fraud cue-based methods) have been proposed to identify phishing websites. The …","url":["https://ieeexplore.ieee.org/abstract/document/9568704/"]}
{"year":"2021","title":"An Exploration of Alignment Concepts to Bridge the Gap between Phrase-based and Neural Machine Translation","authors":["JT Peter"],"snippet":"Page 1. An Exploration of Alignment Concepts to Bridge the Gap between Phrase-based and Neural Machine Translation Von der Fakultät für Mathematik, Informatik und Naturwissenschaften der RWTH Aachen University zur …","url":["https://www-i6.informatik.rwth-aachen.de/publications/download/1175/PeterJan-Thorsten--ExplorationofAlignmentConceptstoBridgetheGapbetweenPhrase-basedNeuralMachineTranslation--2020.pdf"]}
{"year":"2021","title":"An Exploratory Analysis of Multilingual Word-Level Quality Estimation with Cross-Lingual Transformers","authors":["T Ranasinghe, C Orasan, R Mitkov - arXiv preprint arXiv:2106.00143, 2021"],"snippet":"… Our architecture relies on the XLM-R transformer model (Conneau et al., 2020) to derive the representations of the input sentences. XLM-R has been trained on a large-scale multilingual dataset in 104 languages, totalling …","url":["https://arxiv.org/pdf/2106.00143"]}
{"year":"2021","title":"An Exploratory Study on Utilising the Web of Linked Data for Product Data Mining","authors":["Z Zhang, X Song - arXiv preprint arXiv:2109.01411, 2021"],"snippet":"… The Web Data Commons3 (WDC) project extracts such structured data from the CommonCrawl4 as RDF n-quads5, and release them on … to create a very large training dataset for product entity linking using semantic markup data …","url":["https://arxiv.org/pdf/2109.01411"]}
{"year":"2021","title":"An extended analysis of the persistence of persistent identifiers of the scholarly web","authors":["M Klein, L Balakireva - International Journal on Digital Libraries, 2021"],"snippet":"… These findings were confirmed in a large-scale study by Thompson and Jian [22] based on two samples of the web taken from Common Crawl Footnote 3 datasets. The authors were motivated to quantify the use of HTTP DOIs versus URLs of …","url":["https://link.springer.com/article/10.1007/s00799-021-00315-w"]}
{"year":"2021","title":"An Intrinsic and Extrinsic Evaluation of Learned COVID-19 Concepts using Open-Source Word Embedding Sources","authors":["S Parikh, A Davoudi, S Yu, C Giraldo, E Schriver… - medRxiv"],"snippet":"… 8] and GloVe [9] on large corpora of texts including domain-independent texts (eg, internet web pages like Wikipedia and CommonCrawl; social media … Standard GloVe Embeddings Paper Vectors [9] Common Crawl Token 10 …","url":["https://www.medrxiv.org/content/medrxiv/early/2021/01/04/2020.12.29.20249005.full.pdf"]}
{"year":"2021","title":"An Investigation towards Differentially Private Sequence Tagging in a Federated Framework","authors":["A Jana, C Biemann"],"snippet":"… 2The hyperparameter settings to train those models are as follows: epochs- 10, batch size - 32, learning rate - 0.15, optimizer - Stochastic gradient descent (SGD) Common Crawl corpus) from spaCy library3, the dimension of which is 300 …","url":["https://www.inf.uni-hamburg.de/en/inst/ab/lt/publications/2021-janabiemann-privnlp-fed.pdf"]}
{"year":"2021","title":"An Overview on Evaluation Labs and Open Issues in Health-related Credible Information Retrieval","authors":["R Upadhyay, G Pasi, M Viviani - 2021"],"snippet":"… The 2020 Track used a dataset provided by Common Crawl, in particular related to different news collected in the first four months of 2020.4 On … 2https://trec-health-misinfo.github.io/2019.html 3https://lemurproject.org …","url":["http://52.178.216.184/paper31.pdf"]}
{"year":"2021","title":"Analysis and Evaluation of Language Models for Word Sense Disambiguation","authors":["D Loureiro, K Rezaee, MT Pilehvar… - Computational Linguistics, 2021"],"snippet":"Page 1. Analysis and Evaluation of Language Models for Word Sense Disambiguation Daniel Loureiro∗ LIAAD - INESC TEC Department of Computer Science - FCUP University of Porto, Portugal dloureiro@fc.up.pt Kiamehr …","url":["https://direct.mit.edu/coli/article-pdf/doi/10.1162/coli_a_00405/1900170/coli_a_00405.pdf"]}
{"year":"2021","title":"Analysis of Machine Learning and Deep Learning Frameworks for Opinion Mining on Drug Reviews","authors":["F Youbi, N Settouti - The Computer Journal, 2021"],"snippet":"… More precisely, GloVe consists of collecting word co- occurrence statistics in a form of a word co-occurrence matrix, in which its developers have provided pre-embed millions of English tokens obtained from Wikipedia data and common crawl data …","url":["https://academic.oup.com/comjnl/advance-article-abstract/doi/10.1093/comjnl/bxab084/6311550"]}
{"year":"2021","title":"Analyzing Hyperonyms of Stack Overflow Posts","authors":["L Tóth, L Vidács"],"snippet":"… They applied a similar lexico-syntactic pattern-based mining on the dataset obtained from CommonCrawl [17] using a slightly different grammar for NP identification and, therefore, a slightly different set of patterns. Despite the differences …","url":["https://www.researchgate.net/profile/Laszlo-Toth-12/publication/356192289_Analyzing_Hyperonyms_of_Stack_Overflow_Posts/links/61910421d7d1af224bea68e9/Analyzing-Hyperonyms-of-Stack-Overflow-Posts.pdf"]}
{"year":"2021","title":"Analyzing Multimodal Language via Acoustic-and Visual-LSTM with Channel-aware Temporal Convolution Network","authors":["S Mai, S Xing, H Hu - IEEE/ACM Transactions on Audio, Speech, and …, 2021"],"snippet":"Page 1. 2329-9290 (c) 2021 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This …","url":["https://ieeexplore.ieee.org/abstract/document/9387606/"]}
{"year":"2021","title":"Analyzing the Forgetting Problem in Pretrain-Finetuning of Open-domain Dialogue Response Models","authors":["T He, J Liu, K Cho, M Ott, B Liu, J Glass, F Peng - … of the 16th Conference of the …, 2021"],"snippet":"… 4.1 Datasets For pretraining, we use the large-scale CCNEWS data (Bakhtin et al., 2019) which is a de-duplicated subset of the English portion of the CommonCrawl news dataset1 … We tune the 1 http://commoncrawl.org/2016/10/ news-dataset-available Page 5. 1125 …","url":["https://www.aclweb.org/anthology/2021.eacl-main.95.pdf"]}
{"year":"2021","title":"Analyzing transfer learning impact in biomedical cross-lingual named entity recognition and normalization","authors":["RM Rivera-Zavala, P Martínez - BMC Bioinformatics, 2021"],"snippet":"… The FastText-2M [52] pre-trained English word embedding model trained with subword information on Common Crawl using the FastText implementation. Finally, the PubMed and PMC [53] pre-trained English word embedding model, trained on a …","url":["https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-021-04247-9"]}
{"year":"2021","title":"Annotation of Fine-Grained Geographical Entities in German Texts","authors":["J Moreno-Schneider, M Plakidis, G Rehm - 3rd Conference on Language, Data and …, 2021"],"snippet":"… The SpaCy models are trained on Ontonotes 5 and Common Crawl (English; en_core_web_md) and WikiNER and TIGER (German; de_core_news_md). The Stanford models are trained on the CoNLL 2003 data [18]. BERT-NER is trained on WikiNER [11] …","url":["https://drops.dagstuhl.de/opus/volltexte/2021/14547/pdf/OASIcs-LDK-2021-11.pdf"]}
{"year":"2021","title":"Answering questions about insurance supervision with a Neural Machine Translator","authors":["J Glowienke - 2021"],"snippet":"… Conneau et al. [3] pre-train a model based on the RoBERTa architecture to create crosslingual representations. The XLM-R model is pre-trained on a common crawl dataset of 100 languages using the masked multi-lingual language model ap- proach …","url":["https://dke.maastrichtuniversity.nl/jan.niehues/wp-content/uploads/2021/08/Glowienke-Master-thesis.pdf"]}
{"year":"2021","title":"Anticipating Attention: On the Predictability of News Headline Tests","authors":["N Hagar, N Diakopoulos, B DeWilde - Digital Journalism, 2021"],"snippet":"… These embeddings contain 300 dimensions and were trained on English language text from the OntoNotes 5.0 and GloVe Common Crawl corpora. For each headline, we computed the average embedding vector across all tokens. …","url":["https://www.tandfonline.com/doi/abs/10.1080/21670811.2021.1984266"]}
{"year":"2021","title":"Applying and Understanding an Advanced, Novel Deep Learning Approach: A Covid 19, Text Based, Emotions Analysis Study","authors":["J Choudrie, S Patil, K Kotecha, N Matta, I Pappas - Information Systems Frontiers, 2021"],"snippet":"The pandemic COVID 19 has altered individuals' daily lives across the globe. It has led to preventive measures such as physical distancing to be impo.","url":["https://link.springer.com/article/10.1007/s10796-021-10152-6"]}
{"year":"2021","title":"Applying Deep Learning Techniques for Sentiment Analysis to Assess Sustainable Transport","authors":["A Serna Nocedal, A Soroa Echave, R Agerri Gascón - 2021","A Serna, A Soroa, R Agerri - Sustainability, 2021"],"snippet":"… Thus, the multilingual version of BERT [25] was trained for 104 languages. More recently, XLM-RoBERTa [21] distributes a multilingual model which contains 100 languages trained on 2.5 TB of filtered Common Crawl text. To …","url":["https://addi.ehu.eus/bitstream/handle/10810/50497/sustainability-13-02397-v2.pdf?sequence=1&isAllowed=y","https://www.mdpi.com/2071-1050/13/4/2397/pdf"]}
{"year":"2021","title":"Applying Deep Learning Techniques for Sentiment Analysis to Assess Sustainable Transport. Sustainability 2021, 13, 2397","authors":["A Serna, A Soroa, R Agerri - 2021"],"snippet":"… Thus, the multilingual version of BERT [25] was trained for 104 languages. More recently, XLM-RoBERTa [21] distributes a multilingual model which contains 100 languages trained on 2.5 TB of filtered Common Crawl text. To …","url":["https://search.proquest.com/openview/b1ea0637935ea567d5fd68853527c980/1?pq-origsite=gscholar&cbl=2032327"]}
{"year":"2021","title":"AR-LSAT: Investigating Analytical Reasoning of Text","authors":["W Zhong, S Wang, D Tang, Z Xu, D Guo, J Wang, J Yin… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. AR-LSAT: Investigating Analytical Reasoning of Text Wanjun Zhong1∗, Siyuan Wang3∗, Duyu Tang2, Zenan Xu1∗, Daya Guo1∗ Jiahai Wang1, Jian Yin1, Ming Zhou4 and Nan Duan2 1 The School of Data and Computer Science, Sun Yat-sen University …","url":["https://arxiv.org/pdf/2104.06598"]}
{"year":"2021","title":"Arabic Offensive Language Detection in Social Media","authors":["F Husain - 2021"],"snippet":"Page 1. ARABIC OFFENSIVE LANGUAGE DETECTION IN SOCIAL MEDIA by Fatemah Ali Husain A Dissertation Submitted to the Graduate Faculty of George Mason University in Partial Fulfillment of The Requirements for the …","url":["https://search.proquest.com/openview/aefe47a620c621b1c7ed7f95196cf6ba/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"AraCOVID19-SSD: Arabic COVID-19 Sentiment and Sarcasm Detection Dataset","authors":["MS Hadj Ameur - Revue de l'Information Scientifique et Technique, 2023","MSH Ameur, H Aliane - arXiv preprint arXiv:2110.01948, 2021"],"snippet":"… Multilingual BERT (mBERT)6: A BERT-based model [17] pretrained on the first 104 major Wikipedia languages7. • XLM-Roberta 8: A large multi-lingual language model, trained on 2.5TB of filtered Common Crawl data [19]. 4.2.2 Bag-of-Words Models …","url":["https://arxiv.org/pdf/2110.01948","https://www.asjp.cerist.dz/index.php/en/downArticle/134/27/1/220363"]}
{"year":"2021","title":"AraStance: A Multi-Country and Multi-Domain Dataset of Arabic Stance Detection for Fact Checking","authors":["T Alhindi, A Alabdulkarim, A Alshehri, M Abdul-Mageed… - arXiv preprint arXiv …, 2021"],"snippet":"… AraStance and Khoja. This indicates the suitability of the pretraining data of ARBERT that includes Books, Gi- gawords and Common Crawl data primarily from MSA but also a small amount of Egyptian Arabic. Since half of …","url":["https://arxiv.org/pdf/2104.13559"]}
{"year":"2021","title":"ARBERT & MARBERT: Deep Bidirectional Transformers for Arabic","authors":["M Abdul-Mageed, AR Elmadany, EMB Nagoudi - arXiv preprint arXiv:2101.01785, 2020"],"snippet":"… mBERT (Devlin et al., 2019) and XLM-R (Conneau et al., 2020) use a small Arabic text collection from Wikipedia (153M tokens) and CommonCrawl (2.9B … XLM-R (Conneau et al., 2020) is trained on Common Crawl data, hence …","url":["https://arxiv.org/pdf/2101.01785"]}
{"year":"2021","title":"Are Multilingual Models Effective in Code-Switching?","authors":["GI Winata, S Cahyawijaya, Z Liu, Z Lin, A Madotto… - arXiv preprint arXiv …, 2021"],"snippet":"… switching tasks. 2.2.2 XLM-RoBERTa XLM-RoBERTa (XLM-R) (Conneau et al., 2020) is a multilingual language model that is pre-trained on 100 languages using more than two terabytes of filtered CommonCrawl data. Thanks to …","url":["https://arxiv.org/pdf/2103.13309"]}
{"year":"2021","title":"Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? A Comprehensive Assessment for Catalan","authors":["J Armengol-Estapé, CP Carrino, C Rodriguez-Penagos… - arXiv preprint arXiv …, 2021"],"snippet":"… Catalan Government; (2) the Catalan Open Subtitles, a collection of translated movie subtitles (Tiedemann, 2012); (3) the non-shuffled version of the Catalan part of the OSCAR corpus (Suárez et al., 2019), a collection …","url":["https://arxiv.org/pdf/2107.07903"]}
{"year":"2021","title":"Are You Really Complaining? A Multi-task Framework for Complaint Identification, Emotion, and Sentiment Classification","authors":["A Singh, S Saha - International Conference on Document Analysis and …, 2021"],"snippet":"… For deep learning baseline (MT\\(_{\\mathrm{GloVe}}\\)) we also used pre-trained GloVe 13 [16] word embedding which is trained on Common Crawl (840 billion tokens) corpus to get the word embedding representations. 4.4 Results and Discussion …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86331-9_46"]}
{"year":"2021","title":"ARMAN: Pre-training with Semantically Selecting and Reordering of Sentences for Persian Abstractive Summarization","authors":["A Salemi, E Kebriaei, GN Minaei, A Shakery - arXiv preprint arXiv:2109.04098, 2021"],"snippet":"… This corpus contains around 2.8M articles and 1.4B words in all of the articles. CC100 (Conneau et al., 2020; Wenzek et al., 2020) is a monolingual dataset for 100+ languages constructed from Commoncrawl snapshots. This …","url":["https://arxiv.org/pdf/2109.04098"]}
{"year":"2021","title":"As Easy as 1, 2, 3: Behavioural Testing of NMT Systems for Numerical Translation","authors":["J Wang, C Xu, F Guzman, A El-Kishky, BIP Rubinstein… - arXiv preprint arXiv …, 2021"],"snippet":"… Our testing framework facilitates constructing test instances for new domains in the following steps: 1. Obtain a large corpus of text that contains numbers (eg,CommonCrawl); 2. Check if there is a number in the output translation; …","url":["https://arxiv.org/pdf/2107.08357"]}
{"year":"2021","title":"Aspect-based Sentiment Analysis with Graph Convolution over Syntactic Dependencies","authors":["A Zunic, P Corcoran, I Spasic","A Žunić, P Corcoran, I Spasić - Artificial Intelligence in Medicine, 2021"],"snippet":"… individual sentences into dependency graphs. Individual words representing vertices in such graphs were mapped onto their embeddings, which were pretrained 160 on web data from Common Crawl using the GloVe method [31]. Each input …","url":["https://www.researchgate.net/profile/Irena-Spasic/publication/353775596_Aspect-based_Sentiment_Analysis_with_Graph_Convolution_over_Syntactic_Dependencies/links/61112fae169a1a0103ea3e67/Aspect-based-Sentiment-Analysis-with-Graph-Convolution-over-Syntactic-Dependencies.pdf","https://www.sciencedirect.com/science/article/pii/S0933365721001317"]}
{"year":"2021","title":"ASR4REAL: An extended benchmark for speech models","authors":["M Riviere, J Copet, G Synnaeve - arXiv preprint arXiv:2110.08583, 2021"],"snippet":"… even a language model trained on a dataset as big as Common Crawl does not seem to have significant positive effect which reiterates … For all of theses models we used the a 4-gram LM trained on Common Crawl with the decoding parameters …","url":["https://arxiv.org/pdf/2110.08583"]}
{"year":"2021","title":"Assessing reasoning and world knowledge of large language models using questionized counterfactual conditionals","authors":["J Frohberg, F Binder - 2021"],"snippet":"Page 1. Assessing reasoning and world knowledge of large language models using questionized counterfactual conditionals Jörg Frohberg apergo UG Leipzig, Germany j.frohberg@apergo.ai Frank Binder Institute for Applied …","url":["https://openreview.net/pdf?id=i9XYDrUJYyP"]}
{"year":"2021","title":"Assessing the Extent and Types of Hate Speech in Fringe Communities: A Case Study of Alt-Right Communities on 8chan, 4chan, and Reddit","authors":["D Rieger, AS Kümpel, M Wich, T Kiening, G Groh - Social Media+ Society, 2021"],"snippet":"… For this article, the fastText word vectors pre-trained on the English Common Crawl dataset were used because it is trained on web data and thus an appropriate basis (Mikolov, Grave, Bojanowski, Puhrsch, & Joulin, 2019). …","url":["https://journals.sagepub.com/doi/pdf/10.1177/20563051211052906"]}
{"year":"2021","title":"AStitchInLanguageModels: Dataset and Methods for the Exploration of Idiomaticity in Pre-Trained Language Models","authors":["HT Madabushi, E Gow-Smith, C Scarton… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. AStitchInLanguageModels: Dataset and Methods for the Exploration of Idiomaticity in Pre-Trained Language Models Harish Tayyar Madabushi, Edward Gow-Smith, Carolina Scarton and Aline Villavicencio Department …","url":["https://arxiv.org/pdf/2109.04413"]}
{"year":"2021","title":"Attention-based model for predicting question relatedness on Stack Overflow","authors":["J Pei, Z Qin, Y Cong, J Guan - arXiv preprint arXiv:2103.10763, 2021"],"snippet":"… released by Stanford [22]. This word embeddings pre-trained in the Common Crawl corpus, which contains a large amount of data irrelevant to software engineering, may lead to ambiguous results [18]. Therefore, we hope that …","url":["https://arxiv.org/pdf/2103.10763"]}
{"year":"2021","title":"Attention: there is an inconsistency between android permissions and application metadata!","authors":["H Alecakir, B Can, S Sen - International Journal of Information Security"],"snippet":"Since mobile applications make our lives easier, there is a large number of mobile applications customized for our needs in the application markets. While.","url":["https://link.springer.com/article/10.1007/s10207-020-00536-1"]}
{"year":"2021","title":"Attentive Excitation and Aggregation for Bilingual Referring Image Segmentation","authors":["Q Zhou, T Hui, R Wang, H Hu, S Liu - ACM Transactions on Intelligent Systems and …, 2021"],"snippet":"… For English expression, we use GloVe1 pretrained on Common Crawl to embed each word into a 300-d vector. For Chinese expression, existing tools … GloVe word embeddings [34] pretrained on Common Crawl 840B …","url":["https://dl.acm.org/doi/abs/10.1145/3446345"]}
{"year":"2021","title":"Augmenting Poetry Composition with Verse by Verse","authors":["D Uthus, M Voitovich, RJ Mical - arXiv preprint arXiv:2103.17205, 2021"],"snippet":"… TextSETTR was shown to yield better results in transforming sentiment while preserving fluency (important aspects for our work). As described in the TextSETTR paper, we use the model that had been fine-tuned on English Common Crawl data …","url":["https://arxiv.org/pdf/2103.17205"]}
{"year":"2021","title":"Augmenting semantic lexicons using word embeddings and transfer learning","authors":["T Alshaabi, C Van Oort, M Fudolig, MV Arnold… - arXiv preprint arXiv …, 2021"],"snippet":"… words. We then pass the token embeddings to a 300dimensional embedding layer. We initialize the embedding layer with weights trained with subword information on Common Crawl and Wikipedia using FastText [59]. In …","url":["https://arxiv.org/pdf/2109.09010"]}
{"year":"2021","title":"AUGVIC: Exploiting BiText Vicinity for Low-Resource NMT","authors":["T Mohiuddin, MS Bari, S Joty - arXiv preprint arXiv:2106.05141, 2021"],"snippet":"… localization guide, respectively. For some languages, the amount of specific domain monolingual data is limited, where we added additional monolingual data of that language from Common Crawl. Following previous work …","url":["https://arxiv.org/pdf/2106.05141"]}
{"year":"2021","title":"Authorship Weightage Algorithm for Academic publications: A new calculation and ACES webserver for determining expertise","authors":["WL Wu, O Tan, KF Chan, NB Ong, D Gunasegaran… - Methods and Protocols, 2021"],"snippet":"… the back-end server. These word vectors were trained on Common Crawl (https://commoncrawl.org (last accessed on 28 April 2021)) using fastText [17], and are used to map the processed query to its corresponding values …","url":["https://www.mdpi.com/2409-9279/4/2/41/pdf"]}
{"year":"2021","title":"Automated Change Detection in Privacy Policies","authors":["A Adhikari - 2020"],"snippet":"Page 1. University of Denver Digital Commons @ DU Electronic Theses and Dissertations Graduate Studies 2020 Automated Change Detection in Privacy Policies Andrick Adhikari Follow this and additional works at: https://digitalcommons.du.edu/etd …","url":["https://digitalcommons.du.edu/cgi/viewcontent.cgi?article=2702&context=etd"]}
{"year":"2021","title":"Automated essay scoring: A review of the field","authors":["P Lagakis, S Demetriadis - … International Conference on Computer, Information and …, 2021"],"snippet":"… Transformer models make use of those huge datasets of existing general text data, such as Wikipedia Corpus and Common Crawl, to pretrain multilayer neural networks with context-sensitive meaning of, and relations between, words, such as …","url":["https://ieeexplore.ieee.org/abstract/document/9618476/"]}
{"year":"2021","title":"Automated Grading of Exam Responses: An Extensive Classification Benchmark","authors":["A Farazouli, Z Lee, P Papapetrou, U Fors - … Science: 24th International Conference, DS 2021 …","J Ljungman, V Lislevand, J Pavlopoulos, A Farazouli… - International Conference on …, 2021"],"snippet":"… This method proves that training BERT with alternative design choices and with more data, including the CommonCrawl News dataset, … training XLM-R on one hundred languages using CommonCrawl data2, in contrast to previous works such …","url":["https://books.google.de/books?hl=en&lr=lang_en&id=IydHEAAAQBAJ&oi=fnd&pg=PA3&dq=commoncrawl&ots=QIe2sENq0_&sig=LQ1NnDlylvNDV4-vNPAiGJEMZd4","https://link.springer.com/chapter/10.1007/978-3-030-88942-5_1"]}
{"year":"2021","title":"Automated identification of bias inducing words in news articles using linguistic and context-oriented features","authors":["T Spinde, L Rudnitckaia, J Mitrović, F Hamborg… - Information Processing & …, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0306457321000157"]}
{"year":"2021","title":"Automated methods for Question-Answering in Icelandic","authors":["V Snæbjarnarson"],"snippet":"… The source of the data is the open internet, made accessible to those with relatively modest computing resources and disk storage through the targeted use of the Common Crawl datasets that comprise petabytes of data. Prior work has focused on the …","url":["https://vesteinn.is/thesis_150921.pdf"]}
{"year":"2021","title":"Automatic Detection of Fake","authors":["BM Bažík"],"snippet":"… For the data, they created the RealNews dataset, a large corpus of news articles from Common Crawl1. Fake News Detection Using Deep Learning Techniques [11] compared Logistic Regression (LR), Naive Bayes (NB) …","url":["https://is.muni.cz/th/hk1px/Martin_Bazik_master_thesis.pdf"]}
{"year":"2021","title":"Automatic Difficulty Classification of Arabic Sentences","authors":["N Khallaf, S Sharoff - arXiv preprint arXiv:2103.04386, 2021"],"snippet":"… corpus (Common Crawl and Wikipedia for ArabicBERT vs Common Crawl XML-R vs Wikipedia for BERT, AraBert and UCS) used to train the Arabic … The corpus will be classified on the ba- sis of how difficult the sentences are …","url":["https://arxiv.org/pdf/2103.04386"]}
{"year":"2021","title":"Automatic Fully-Contextualized Recommendation Extraction from Radiology Reports","authors":["J Steinkamp, C Chambers, D Lalevic, T Cook - Journal of Digital Imaging, 2021"],"snippet":"… We evaluated a simple long short-term memory (LSTM) architecture [12] on the task. We used a combination of custom-trained fastText vectors, trained on our institution's entire repository of radiology reports, with Global …","url":["https://link.springer.com/article/10.1007/s10278-021-00423-8"]}
{"year":"2021","title":"Automatic Generic Web Information Extraction at Scale","authors":["M Aljabary - 2021"],"snippet":"Page 1. 1 Page 2. 2 Automatic Generic Web Information Extraction at Scale Master Thesis Computer Science, Data Science and Technology University of Twente. Enschede, The Netherlands An attempt to bring some structure …","url":["http://essay.utwente.nl/86153/1/Aljabary_MA_EEMCS.pdf"]}
{"year":"2021","title":"Automatic Sexism Detection with Multilingual Transformer Models","authors":["S Mina, B Jaqueline, L Daria, S Djordje, K Armin… - arXiv preprint arXiv …, 2021"],"snippet":"… XLM-R is a multilingual model trained on 100 languages, similar to mBERT. Unlike the latter, XLM-R is not trained on Wikipedia data but on monolingual CommonCrawl data. The model shows improved cross-lingual language …","url":["https://arxiv.org/pdf/2106.04908"]}
{"year":"2021","title":"Automatic Stress Detection from Facial Videos","authors":["EM de Oca - 2021"],"snippet":"… , leading to the development of pretrained systems such as BERT(Bidirectional Encoder Representations from Transformers) and GPT (Generative Pre-trained Transformer), which were trained with large language datasets, such as Wikipedia …","url":["https://www.eduardomontesdeoca.com/s/ADSS-Project.pdf"]}
{"year":"2021","title":"Automatically Detecting Cyberbullying Comments on Online Game Forums","authors":["HHP Vo, HT Tran, ST Luu - arXiv preprint arXiv:2106.01598, 2021"],"snippet":"… on Wikipedia and Gigaword corpora. The fastText 3 is trained on Common Crawl and Wikipedia datasets using CBOW with position-weight in 300 dimensions with 5-grams features. D. Traditional machine learning models Logistic …","url":["https://arxiv.org/pdf/2106.01598"]}
{"year":"2021","title":"Autonomous Writing Futures","authors":["AH Duin, I Pedersen - Writing Futures: Collaborative, Algorithmic …, 2021"],"snippet":"… GPT-3 is 175 billion parameters. GPT-3 is trained on the Common Crawl data set, a corpus of almost a trillion words of texts scraped from the Web. “The dataset and model size are about two orders of magnitude larger than those used for GPT-2,” the authors write …","url":["https://link.springer.com/chapter/10.1007/978-3-030-70928-0_4"]}
{"year":"2021","title":"Auxiliary Bi-Level Graph Representation for Cross-Modal Image-Text Retrieval","authors":["X Zhong, Z Yang, M Ye, W Huang, J Yuan, CW Lin - 2021 IEEE International …, 2021"],"snippet":"… The scene graph features Soi and Srij are transformed by a learnable embedding layer which is initialized by GloVe [18] pre-trained on the Common-Crawl dataset, and maps Ioi and Irij into a vector of same dimension: Soi = WoIoi , Srij = WrIrij , (1) …","url":["https://ieeexplore.ieee.org/abstract/document/9428380/"]}
{"year":"2021","title":"Auxiliary Learning for Relation Extraction","authors":["S Lyu, J Cheng, X Wu, L Cui, H Chen, C Miao - IEEE Transactions on Emerging …, 2020"],"snippet":"… 7https://catalog.ldc.upenn.edu/LDC2018T24 8http://semeval2.fbk.eu/semeval2. php?location=data 9Following previous work, we choose GloVe word vectors with 300 dimensions (Common Crawl) https://nlp.stanford.edu/projects/glove …","url":["https://ieeexplore.ieee.org/abstract/document/9296307/"]}
{"year":"2021","title":"Background Knowledge in Schema Matching: Strategy vs. Data","authors":["J Portisch, M Hladik, H Paulheim - arXiv preprint arXiv:2107.00001, 2021"],"snippet":"… used. WebIsALOD is a large hypernymy graph based on the WebIsA database [37]. The latter is a dataset which consists of hypernymy relations extracted from the Common Crawl, a large set of crawled Web pages. The extraction …","url":["https://arxiv.org/pdf/2107.00001"]}
{"year":"2021","title":"Bambara Language Dataset for Sentiment Analysis","authors":["M Diallo, C Fourati, H Haddad - arXiv preprint arXiv:2108.02524, 2021"],"snippet":"… In this paper, we present the first common-crawl-based Bambara dialectal dataset dedicated for Sentiment Analysis, available freely for Natural Language Processing research purposes … Bambara V1 dataset represents …","url":["https://arxiv.org/pdf/2108.02524"]}
{"year":"2021","title":"Bandits Don't Follow Rules: Balancing Multi-Facet Machine Translation with Multi-Armed Bandits","authors":["J Kreutzer, D Vilar, A Sokolov - arXiv preprint arXiv:2110.06997, 2021"],"snippet":"Training data for machine translation (MT) is often sourced from a multitude of large corpora that are multi-faceted in nature, eg containing contents from multiple domains or different levels of quality or complexity. Naturally, these facets do not …","url":["https://arxiv.org/pdf/2110.06997"]}
{"year":"2021","title":"BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese","authors":["NL Tran, DM Le, DQ Nguyen - arXiv preprint arXiv:2109.09701, 2021"],"snippet":"… rization. Here, mBART is pre-trained on a Common Crawl dataset of 25 languages, which contains 137 GB of syllablelevel Vietnamese texts. We employ the single-document summarization dataset VNDS (Nguyen et al. 2019 …","url":["https://arxiv.org/pdf/2109.09701"]}
{"year":"2021","title":"belabBERT: a Dutch RoBERTa-based language model applied to psychiatric classification","authors":["J Wouts, J de Boer, A Voppel, S Brederoo… - arXiv preprint arXiv …, 2021"],"snippet":"… 3.1.1. Pre-training For the pre-training of belabBERT we used the OSCAR corpus which consists of a set of monolingual corpora extracted from Common Crawl snapshots … belabBERT Common Crawl Dutch (non-shuffled) BytePairEncoding 95.92 ∗ …","url":["https://arxiv.org/pdf/2106.01091"]}
{"year":"2021","title":"Benchmarking Differential Privacy and Federated Learning for BERT Models","authors":["P Basu, TS Roy, R Naidu, Z Muftuoglu, S Singh… - arXiv preprint arXiv …, 2021"],"snippet":"… It uses 160 GB of text for pre-training, including 16GB of Books Corpus and English Wikipedia used in BERT. The additional data included CommonCrawl News dataset, Web text corpus and Stories from Common Crawl. For …","url":["https://arxiv.org/pdf/2106.13973"]}
{"year":"2021","title":"BERT, mBERT, or BiBERT? A Study on Contextualized Embeddings for Neural Machine Translation","authors":["H Xu, B Van Durme, K Murray - arXiv preprint arXiv:2109.04588, 2021"],"snippet":"… on 145G German text data portion of OSCAR (Or- tiz Suárez et al., 2020), a huge multilingual corpus extracted from Common Crawl … Vaswani et al., 2017) masked language model trained on 100 languages, using more than …","url":["https://arxiv.org/pdf/2109.04588"]}
{"year":"2021","title":"BERT: A Review of Applications in Natural Language Processing and Understanding","authors":["MV Koroteev - arXiv preprint arXiv:2103.11943, 2021"],"snippet":"Page 1. BERT: A Review of Applications in Natural Language Processing and Understanding Koroteev MV, Financial University under the government of the Russian Federation, Moscow, Russia mvkoroteev@fa.ru Abstract: In …","url":["https://arxiv.org/pdf/2103.11943"]}
{"year":"2021","title":"Bertinho: Galician BERT Representations","authors":["D Vilares, M Garcia, C Gómez-Rodríguez - arXiv preprint arXiv:2103.13799, 2021"],"snippet":"Page 1. Bertinho: Galician BERT Representations Bertinho: Representaciones BERT para el gallego David Vilares,1 Marcos Garcia,2 Carlos Gómez-Rodr´ıguez 1 1Universidade da Coru˜na, CITIC, Galicia, Spain 2CiTIUS, Universidade …","url":["https://arxiv.org/pdf/2103.13799"]}
{"year":"2021","title":"Better Neural Machine Translation by Extracting Linguistic Information from BERT","authors":["HS Shavarani, A Sarkar - arXiv preprint arXiv:2104.02831, 2021"],"snippet":"… clack, clack.”). 9Europarl+CommonCrawl+NewsCommentary https://www.statmt. org/wmt14/translation-task.html, please note that in the later years this training set remained the same, but ParaCrawl data was added to it. We …","url":["https://arxiv.org/pdf/2104.02831"]}
{"year":"2021","title":"Beyond Noise: Mitigating the Impact of Fine-grained Semantic Divergences on Neural Machine Translation","authors":["E Briakou, M Carpuat - arXiv preprint arXiv:2105.15087, 2021"],"snippet":"Page 1. Beyond Noise: Mitigating the Impact of Fine-grained Semantic Divergences on Neural Machine Translation Eleftheria Briakou and Marine Carpuat Department of Computer Science University of Maryland College Park …","url":["https://arxiv.org/pdf/2105.15087"]}
{"year":"2021","title":"Beyond the English Web: Zero-Shot Cross-Lingual and Lightweight Monolingual Classification of Registers","authors":["L Repo, V Skantsi, S Rönnqvist, S Hellström… - arXiv preprint arXiv …, 2021"],"snippet":"… FreCORE and SweCORE are random samples of the 2017 CoNLL datasets (Ginter et al., 2017) originally drawn from Common Crawl … XLM-R is trained on 2.5TB of filtered Common Crawl (Wenzek et al., 2020) data comprising …","url":["https://arxiv.org/pdf/2102.07396"]}
{"year":"2021","title":"Bias Silhouette Analysis: Towards Assessing the Quality of Bias Metrics for Word Embedding Models","authors":["M Spliethöver, H Wachsmuth"],"snippet":"… Word Embedding Models. As biased and unbiased models, we use GloVe CommonCrawl [Pennington et al., 2014] trained on 840 billion English tokens and the English ConceptNet Numberbatch 19.08 [Speer et al., 2017] (referred to as NBatch below), respectively …","url":["https://www.ijcai.org/proceedings/2021/0077.pdf"]}
{"year":"2021","title":"Bidirectional Language Modeling: A Systematic Literature Review","authors":["M Shah Jahan, HU Khan, S Akbar, M Umar Farooq… - Scientific Programming, 2021"],"snippet":"Page 1. Review Article Bidirectional Language Modeling: A Systematic Literature Review Muhammad Shah Jahan ,1 Habib Ullah Khan ,2 Shahzad Akbar ,3 Muhammad Umar Farooq ,1 Sarah Gul ,4 and Anam Amjad 1 1Department …","url":["https://www.hindawi.com/journals/sp/2021/6641832/"]}
{"year":"2021","title":"Bilingual Lexical Induction for Sinhala-English using Cross Lingual Embedding Spaces","authors":["A Liyanage, S Ranathunga, S Jayasena - 2021 Moratuwa Engineering Research …, 2021"],"snippet":"… Using pre-trained fastText embeddings trained on Wikipedia and Common crawl data using two different evaluation dictionaries as a preliminary experiment to identify the performance of embeddings created from non-comparable corpora …","url":["https://ieeexplore.ieee.org/abstract/document/9525667/"]}
{"year":"2021","title":"Bilingual Lexicon Induction via Unsupervised Bitext Construction and Word Alignment","authors":["H Shi, L Zettlemoyer, SI Wang - arXiv preprint arXiv:2101.00148"],"snippet":"Page 1. Bilingual Lexicon Induction via Unsupervised Bitext Construction and Word Alignment Haoyue Shi ∗ TTI-Chicago freda@ttic.edu Luke Zettlemoyer University of Washington Facebook AI Research lsz@fb.com …","url":["https://arxiv.org/pdf/2101.00148"]}
{"year":"2021","title":"BitextEdit: Automatic Bitext Editing for Improved Low-Resource Machine Translation","authors":["E Briakou, SI Wang, L Zettlemoyer, M Ghazvininejad - arXiv preprint arXiv …, 2021"],"snippet":"Mined bitexts can contain imperfect translations that yield unreliable training signals for Neural Machine Translation (NMT). While filtering such pairs out is known to improve final model quality, we argue that it is suboptimal in low-resource conditions …","url":["https://arxiv.org/pdf/2111.06787"]}
{"year":"2021","title":"Blank spots, critical information needs and local journalism fund-ing","authors":["S Bisiani"],"snippet":"Abstract A global business model crisis in journalism, fuelled by loss in advertising revenue, challenges the survival of local news production. In Sweden, it has led to the closure of several newspapers across the country, and the concentration of …","url":["http://compscjournalism.org/projects/simona/projects/Master_Thesis_Simona_Bisiani.pdf"]}
{"year":"2021","title":"Book genre and author's gender recognition based on titles","authors":["A Pawłowski, E Herden, T Walkowiak - … and Text: Data, models, information and …, 2021"],"snippet":""}
{"year":"2021","title":"BOSS: Bandwidth-Optimized Search Accelerator for Storage-Class Memory","authors":["J Heo, SY Lee, S Min, Y Park, SJ Jung, TJ Ham…"],"snippet":"Page 1. BOSS: Bandwidth-Optimized Search Accelerator for Storage-Class Memory Jun Heo, Seung Yul Lee, Sunhong Min, Yeonhong Park, Sung Jun Jung, Tae Jun Ham, Jae W. Lee Seoul National University {j.heo, triomphant1 …","url":["https://conferences.computer.org/iscapub/pdfs/ISCA2021-4ghucdBnCWYB7ES2Pe4YdT/333300a279/333300a279.pdf"]}
{"year":"2021","title":"Bottom-Up Shift and Reasoning for Referring Image Segmentation","authors":["S Yang, M Xia, G Li, HY Zhou, Y Yu - Proceedings of the IEEE/CVF Conference on …, 2021"],"snippet":"Page 1. Bottom-Up Shift and Reasoning for Referring Image Segmentation Sibei Yang1∗† Meng Xia2∗ Guanbin Li2 Hong-Yu Zhou3 Yizhou Yu3,4† 1ShanghaiTech University 2Sun Yat-sen University 3The University of Hong Kong 4Deepwise AI Lab Abstract …","url":["https://openaccess.thecvf.com/content/CVPR2021/papers/Yang_Bottom-Up_Shift_and_Reasoning_for_Referring_Image_Segmentation_CVPR_2021_paper.pdf"]}
{"year":"2021","title":"bradleypallen/keras-quora-question-pairs","authors":["ONDJF Mar, AMJJA Sep, OSMTW Thu, F Sat"],"snippet":"… Model, Source of Word Embeddings, Accuracy. \"BiMPM model\" [5], GloVe Common Crawl (840B tokens, 300D), 0.88 … \"Decomposable attention\" [6], \"Quora's text corpus\", 0.86. \"LDC\" [5], GloVe Common Crawl (840B tokens, 300D), 0.86 …","url":["https://giters.com/bradleypallen/keras-quora-question-pairs?amp=1"]}
{"year":"2021","title":"Building a File Observatory for Secure Parser Development","authors":["T Allison, W Burke, C Mattmann, A Mensikova…"],"snippet":"… 3196–3200. [Online]. Available: http://www.lrec-conf.org/proceedings/lrec2012/pdf/534 Paper.pdf [10] “Common Crawl,” https://commoncrawl.org. [11] P. Wyatt, “Stressful PDF corpus grows!” https://www.pdfa.org/ stressful-pdf-corpus-grows/, November 2020.","url":["https://langsec.org/spw21/papers/Allison_LangSec21.pdf"]}
{"year":"2021","title":"Building a Question and Answer System for News Domain","authors":["S Basu, A Gaddala, P Chetan, G Tiwari, N Darapaneni… - arXiv preprint arXiv …, 2021"],"snippet":"… We have used two approaches for building the Embedding Layers for the models 3. GloVe Embedding: we used the 300 Dimension Common Crawl for the English language 4. Universal Sentence Encoder: we used the 512 …","url":["https://arxiv.org/pdf/2105.05744"]}
{"year":"2021","title":"Building Accountable Natural Language Processing Models: on Social Bias Detection and Mitigation","authors":["J Zhao - 2021"],"snippet":"Natural Language Processing (NLP) plays an important role in many applications, including resume filtering, text analysis, and information retrieval. Despite the remarkable accuracy enabled by the advances of machine learning methods, recent …","url":["https://escholarship.org/content/qt0441n1tt/qt0441n1tt.pdf"]}
{"year":"2021","title":"But how robust is RoBERTa actually?: A Benchmark of SOTA Transformer Networks for Sexual Harassment Detection on Twitter","authors":["P Basu, TS Roy, A Singhal - 2021 Fifth International Conference on I-SMAC (IoT in …, 2021"],"snippet":"Harassment, which is of sexual/physical in nature, is defined as any unwanted sexual misconduct, including the unwarranted and ill-suited promise of benefit in exchange for sexual indulgence. It also includes a span of actions from verbal …","url":["https://ieeexplore.ieee.org/abstract/document/9640861/"]}
{"year":"2021","title":"Can I Take Your Subdomain? Exploring Same-Site Attacks in the Modern Web","authors":["MSMTL Veronese, SCM Maffei"],"snippet":"Page 1. Can I Take Your Subdomain? Exploring Same-Site Attacks in the Modern Web Marco Squarcina1 Mauro Tempesta1 Lorenzo Veronese1 Stefano Calzavara2 Matteo Maffei1 1 TU Wien 2 Università Ca' Foscari Venezia & OWASP …","url":["https://minimalblue.com/data/papers/USENIX21_can_i_take_your_subdomain.pdf"]}
{"year":"2021","title":"Can Language Models Encode Perceptual Structure Without Grounding? A Case Study in Color","authors":["M Abdou, A Kulmizev, D Hershcovich, S Frank… - arXiv preprint arXiv …, 2021"],"snippet":"… Word-type FastText embeddings trained on Common Crawl (Bojanowski et al., 2017) … These S contexts are either randomly sampled from common crawl (RC), or deterministically generated to allow for control over contextual variation (CC) …","url":["https://arxiv.org/pdf/2109.06129"]}
{"year":"2021","title":"Can Small and Synthetic Benchmarks Drive Modeling Innovation? A Retrospective Study of Question Answering Modeling Approaches","authors":["NF Liu, T Lee, R Jia, P Liang - arXiv preprint arXiv:2102.01065, 2021"],"snippet":"Page 1. Can Small and Synthetic Benchmarks Drive Modeling Innovation? A Retrospective Study of Question Answering Modeling Approaches Nelson F. Liu Tony Lee Robin Jia Percy Liang Computer Science Department, Stanford …","url":["https://arxiv.org/pdf/2102.01065"]}
{"year":"2021","title":"CausalBERT: Injecting Causal Knowledge Into Pre-trained Models with Minimal Supervision","authors":["Z Li, X Ding, K Liao, T Liu, B Qin - arXiv preprint arXiv:2107.09852, 2021"],"snippet":"… ambiguity and precise causal patterns to extract word level causeeffect pairs from the preprocessed English Common Crawl corpus (5.14 … (2016) for creating a causal lexical knowledge base, we reproduce a variant of their …","url":["https://arxiv.org/pdf/2107.09852"]}
{"year":"2021","title":"CCQA: A New Web-Scale Question Answering Dataset for Model Pre-Training","authors":["P Huber, A Aghajanyan, B Oğuz, D Okhonko, W Yih… - arXiv preprint arXiv …, 2021"],"snippet":"… Consequently, we propose a novel QA dataset based on the Common Crawl project in this paper. Using the readily available schema.org annotation, we extract around 130 million multilingual question-answer pairs, including about 60 million …","url":["https://arxiv.org/pdf/2110.07731"]}
{"year":"2021","title":"CDA: a Cost Efficient Content-based Multilingual Web Document Aligner","authors":["T Vu, AA AI, A Moschitti - 2021"],"snippet":"… CommonCrawl Sextet Previous datasets share the same domains that are heavily biased toward French content (see Table 3). We leverage a monthly crawl from CommonCrawl, specifically … Table 4: Parallel English tokens …","url":["https://assets.amazon.science/01/69/5f786b844c08a079eda7e6437c16/cda-a-cost-efficient-content-based-multilingual-web-document-aligner.pdf"]}
{"year":"2021","title":"Censorship of Online Encyclopedias: Implications for NLP Models","authors":["E Yang, ME Roberts - arXiv preprint arXiv:2101.09294, 2021"],"snippet":"… Word embeddings are also useful because they can be pre-trained on large corpuses of text like Wikipedia or Common Crawl, and these pre-trained embeddings can then be used as an initial layer in applications that may have less training data …","url":["https://arxiv.org/pdf/2101.09294"]}
{"year":"2021","title":"Challenges for cognitive decoding using deep learning methods","authors":["AW Thomas, C Ré, RA Poldrack - arXiv preprint arXiv:2108.06896, 2021"],"snippet":"… learning in 251 the target domain. Transfer learning has been especially successful in CV and NLP, where large 252 publicly available datasets exist (eg, [72,73] and http://www.commoncrawl.org). Here, DL 253 models are first …","url":["https://arxiv.org/pdf/2108.06896"]}
{"year":"2021","title":"Changing the World by Changing the Data","authors":["A Rogers - arXiv preprint arXiv:2105.13947, 2021"],"snippet":"… The use of uncontrolled samples (like the Common-Crawl-based corpora) would have to be justified by arguing either that the above types of bias can be safely ignored, or that the benefits outweigh the risks. 2.2.3 Might not be the best approach …","url":["https://arxiv.org/pdf/2105.13947"]}
{"year":"2021","title":"Characterizing and addressing the issue of oversmoothing in neural autoregressive sequence modeling","authors":["I Kulikov, M Eremeev, K Cho - arXiv preprint arXiv:2112.08914, 2021"],"snippet":"… We use the subset of WMT’19 training set consisting of news commentary v12 and common crawl resulting in slightly more than 1M and 2M training sentence pairs for Ru→En and De↔En pairs, respectively. We fine-tuned single model checkpoints …","url":["https://arxiv.org/pdf/2112.08914"]}
{"year":"2021","title":"Characterizing Network Infrastructure Using the Domain Name System","authors":["P Kintis - 2020"],"snippet":"Page 1. CHARACTERIZING NETWORK INFRASTRUCTURE USING THE DOMAIN NAME SYSTEM A Dissertation Presented to The Academic Faculty By Panagiotis Kintis In Partial Fulfillment of the Requirements for the Degree …","url":["https://smartech.gatech.edu/bitstream/handle/1853/64165/KINTIS-DISSERTATION-2020.pdf"]}
{"year":"2021","title":"Charformer: Fast Character Transformers via Gradient-based Subword Tokenization","authors":["Y Tay, VQ Tran, S Ruder, J Gupta, HW Chung, D Bahri… - arXiv preprint arXiv …, 2021"],"snippet":"… In addition, we compare to the byte-level models from §3.1, which we pre-train on multilingual data. Setup We pre-train CHARFORMER as well as the Byte-level T5 and Byte-level T5+LASC baselines on multilingual …","url":["https://arxiv.org/pdf/2106.12672"]}
{"year":"2021","title":"ChineseBERT: Chinese Pretraining Enhanced by Glyph and Pinyin Information","authors":["Z Sun, X Li, X Sun, Y Meng, X Ao, Q He, F Wu, J Li - arXiv preprint arXiv:2106.16038, 2021"],"snippet":"… We collected our pretraining data from CommonCrawl5. After pre-processing (such as removing the data with too much … ERNIE BERT-wwm MacBERT ChineseBERT Data Source Heterogeneous Wikipedia Heterogeneous …","url":["https://arxiv.org/pdf/2106.16038"]}
{"year":"2021","title":"Claim Detection in Biomedical Twitter Posts","authors":["A Wührl, R Klinger - arXiv preprint arXiv:2104.11639, 2021"],"snippet":"… The first three models (NB, LG, BiLSTM) use 50-dimensional FastText (Bojanowski et al., 2017) embeddings trained on the Common Crawl corpus (600 billion tokens) as input6. NB. We use a (Gaussian) naive Bayes with …","url":["https://arxiv.org/pdf/2104.11639"]}
{"year":"2021","title":"Classification of Emotions Based on Text and Qualitative Variables","authors":["J Dobša, D Šebalj, D Bužić"],"snippet":"… Experiments were done both with Common Crawl GloVe pretrained vectors with the dimensionality of 300, and without pretrained vectors. Fifteen percent of learning samples were used for validation. We constructed six neural networks models: CNN …","url":["https://www.researchgate.net/profile/Jasminka-Dobsa/publication/355461190_Classification_of_Emotions_Based_on_Text_and_Qualitative_Variables/links/61716c97750da711ac647d77/Classification-of-Emotions-Based-on-Text-and-Qualitative-Variables.pdf"]}
{"year":"2021","title":"Classification of Horror Stories from Reddit","authors":["D Zhou, C Kim, S Gatiganti"],"snippet":"… We further hypothesize that performance would still increase a little if we used the larger pre-trained vectors such as Common Crawl or Twitter sets, but they come with increased download sizes (>1 GB) and increased training time …","url":["http://cs229.stanford.edu/proj2021spr/report2/82008167.pdf"]}
{"year":"2021","title":"Classification of Texts Using a Vocabulary of Synonyms","authors":["A Giliazova - 2021 14th International Conference Management of …, 2021"],"snippet":"… This is a Transformer-based masked language model trained on one hundred languages, including Russian language, using more than two terabytes of filtered CommonCrawl data. The XLM-R model significantly outperforms multilingual BERT (mBERT) …","url":["https://ieeexplore.ieee.org/abstract/document/9600131/"]}
{"year":"2021","title":"CLASSIFICATION OF TWEETS USING MULTIPLE THRESHOLDS WITH SELF-CORRECTION AND WEIGHTED CONDITIONAL","authors":["TN Ahmad - 2020"],"snippet":"Page 1. CLASSIFICATION OF TWEETS USING MULTIPLE THRESHOLDS WITH SELF-CORRECTION AND WEIGHTED CONDITIONAL PROBABILITIES A thesis submitted to The University of Manchester for the degree of Doctor of Philosophy …","url":["https://www.research.manchester.ac.uk/portal/files/188959099/FULL_TEXT.PDF"]}
{"year":"2021","title":"Classification-based Quality Estimation: Small and Efficient Models for Real-world Applications","authors":["S Sun, A El-Kishky, V Chaudhary, J Cross, F Guzmán… - arXiv preprint arXiv …, 2021"],"snippet":"… Current state of the art QE systems (Fomicheva et al., 2020b; Ranasinghe et al., 2020a; Sun et al., 2020). are built on XLM-R (Conneau et al., 2019), a contextualized language model pre-trained on more than 2 terabytes of …","url":["https://arxiv.org/pdf/2109.08627"]}
{"year":"2021","title":"Classifying Fake and Real Neurally Generated News","authors":["A Govindaraju, J Griffith - 2021 Swedish Workshop on Data Science (SweDS), 2021"],"snippet":"… In order to train and test the model, 3 datasets have been created: One containing real news extracted from a common crawl; the second comprises a neural fake news dataset generated using language modelling techniques; the third comprises a …","url":["https://ieeexplore.ieee.org/abstract/document/9638268/"]}
{"year":"2021","title":"CLEF eHealth Evaluation Lab 2021","authors":["L Kelly, LA Alemany, N Brew-Sam, V Cotik, D Filippo…"],"snippet":"… This collection consists of Web pages acquired from Common Crawl,14 which is augmented with additional pages collected from a number of known reliable health Websites and other known unreliable health Websites [9]. The topics …","url":["https://www.researchgate.net/profile/Marco-Viviani/publication/350569762_CLEF_eHealth_Evaluation_Lab_2021/links/6073f32e92851c8a7bbea835/CLEF-eHealth-Evaluation-Lab-2021.pdf"]}
{"year":"2021","title":"Click This, Not That: Extending Web Authentication with Deception","authors":["T Barron, J So, N Nikiforakis - Proceedings of the 2021 ACM Asia Conference on …, 2021"],"snippet":"… after creation. References. 2020. Common Crawl. https://commoncrawl.org/the-data/ get-started/Google Scholar Google Scholar; 2020. Mouseflow: Session Replay, Heatmaps, Funnels, Forms & User Feedback. https://mouseflow …","url":["https://dl.acm.org/doi/abs/10.1145/3433210.3453088"]}
{"year":"2021","title":"ClimateBert: A Pretrained Language Model for Climate-Related Text","authors":["N Webersinke, M Kraus, JA Bingler, M Leippold - arXiv preprint arXiv:2110.12010, 2021"],"snippet":"… 2019), and a subset of CommonCrawl that is said to resemble the storylike style of WINOGRAD schemas (Trinh and Le, 2019). While these sources are valuable to build a model working on general language, it has been shown that domain-specific …","url":["https://arxiv.org/pdf/2110.12010"]}
{"year":"2021","title":"CLIP2StyleGAN: Unsupervised Extraction of StyleGAN Edit Directions","authors":["R Abdal, P Zhu, J Femiani, NJ Mitra, P Wonka - arXiv preprint arXiv:2112.05219, 2021"],"snippet":"… The CLIP image encoder [34] is trained on the common-crawl dataset, an internet-scale set of images that encompasses a broad range of visual concepts. However, a typical high-quality GAN would be trained on a more specific set of images, for example …","url":["https://arxiv.org/pdf/2112.05219"]}
{"year":"2021","title":"Cluster analysis of agricultural household production of self-employed","authors":["AV Plotnikov - IOP Conference Series: Earth and Environmental …, 2021"],"snippet":"… To train this model, we used a sample of Russian-language documents from the CommonCrawl dump, balanced by geography, compiled by Jonathan Dunn and Ben Adams; the corpus Size is 2.1 billion words. Page 5. AGRITECH-IV-2020 IOP Conf …","url":["https://iopscience.iop.org/article/10.1088/1755-1315/677/2/022080/pdf"]}
{"year":"2021","title":"Cluster-Based Antiphishing (CAP) Model for Smart Phones","authors":["M Faisal, S Abed - Scientific Programming"],"snippet":"… latest techniques tested on UCI datasets. 4.4.2. Dataset Taken from Mendeley. Source Phishing web page: Phish Tank, Legitimate web page source: Alexa, Common Crawl (1) Dataset Information. In this scenario, the dataset …","url":["https://www.hindawi.com/journals/sp/2021/9957323/"]}
{"year":"2021","title":"Code-Mixing on Sesame Street: Dawn of the Adversarial Polyglots","authors":["S Tan, S Joty - arXiv preprint arXiv:2103.09593, 2021"],"snippet":"… However, the latter trend is replicated for BUMBLEBEE if we remove this constraint (Table 14 in Appendix G). A possible explanation is that XLM-R and Unicoder were trained on monolingual CommonCrawl (CC) data, while …","url":["https://arxiv.org/pdf/2103.09593"]}
{"year":"2021","title":"CoDesc: A Large Code–Description Parallel Dataset","authors":["M Hasan, T Muttaqueen, A Al Ishtiaq, KS Mehrab…"],"snippet":"… 9052–9065, Online. Association for Computational Linguistics. CommonCrawl Common crawl. https:// commoncrawl.org/. Accessed: 2021-01-31. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019 …","url":["http://masumhasan.net/files/CoDesc.pdf"]}
{"year":"2021","title":"Combining Natural Language Processing and Machine Learning for Profiling and Fake News Detection","authors":["A Bondielli"],"snippet":"Page 1. PHD PROGRAM IN SMART COMPUTING DIPARTIMENTO DI INGEGNERIA DELL'INFORMAZIONE (DINFO) Combining Natural Language Processing and Machine Learning for Profiling and Fake News Detection Alessandro Bondielli …","url":["https://flore.unifi.it/bitstream/2158/1244287/1/PhDThesis_AlessandroBondielli.pdf"]}
{"year":"2021","title":"Combining Pre-trained Word Embeddings and Linguistic Features for Sequential Metaphor Identification","authors":["R Mao, C Lin, F Guerin - arXiv preprint arXiv:2104.03285, 2021"],"snippet":"… For instance, GloVe was trained on Common Crawl2, from billions of web pages (840 billion tokens); ELMo was trained on WMT 2011 News Crawl data3 (800 million tokens); BERT was trained on Wikipedia4 (2.5 billion tokens) …","url":["https://arxiv.org/pdf/2104.03285"]}
{"year":"2021","title":"Combining word embeddings as a tool for subject identification","authors":["A Hamm - Wissensaustauschworkshop\" Maschinelles Lernen VII\", 2021"],"snippet":"This talk shows ungoing work aiming at finding subject matter relations between text documents and word clouds. A number of increasingly successful semantic word embedding procedures - learning semantic relations from contextual distributions …","url":["https://elib.dlr.de/147623/1/Combining%20word%20embeddings.pdf"]}
{"year":"2021","title":"Comparative Analysis of Bengali Stop Word Detection Using Different Approaches","authors":["RJ Rupa, JF Sohana, M Rahman - … on Automation, Control and Mechatronics for …, 2021"],"snippet":"… In this paper, the pretrained Bengali FastText CBOW model is utilized to produce word vectors trained on the common crawl and Wikipedia [27] and both logistic regression and support vector machine classifiers acquire a performance score of 86%. TABLE XII …","url":["https://ieeexplore.ieee.org/abstract/document/9528279/"]}
{"year":"2021","title":"Comparative Analysis of Different Transformer Based Architectures Used in Sentiment Analysis","authors":["K Pipalia, R Bhadja, M Shukla - 2020 9th International Conference System Modeling …, 2020"],"snippet":"… Distill BERT Base:66 BookCorpus wiki BERT Distillation T5 Base:220 large:770 Colossal Clean Crawled Corpus (C4) Text Infilling XLNet Base:~110 Large:~340 BookCorpus Wiki, Giga5 ClueWeb, Common Crawl …","url":["https://ieeexplore.ieee.org/abstract/document/9337081/"]}
{"year":"2021","title":"Comparing Apples and Oranges: Human and Computer Clustered Affinity Diagrams Under the Microscope","authors":["P Borlinghaus, S Huber - 26th International Conference on Intelligent User …, 2021"],"snippet":"… training corpora WSD no OOV LSI [9] − NMF [18] − LDA [5] − GloVe [26] Wiki word2vec [22] Google News corpus doc2vec [17] 900k sentences from qualitative survey ◦ fastText [6] Common Crawl, Wiki • … FastText was trained on Common Crawl and Wikipedia corpus …","url":["https://dl.acm.org/doi/abs/10.1145/3397481.3450674"]}
{"year":"2021","title":"Comparing Contextualised Embeddings for Predicting the (Graded) Effect of Context in Word Similarity","authors":["JM Albers - 2021"],"snippet":"… As data set XLM-RoBERTa uses CommonCrawl instead of Wikipedia, which provides limited scale for low resource languages. 4 Page 5 … The CommonCrawl data set is designed to be more diverse than other data sets, which mainly use Wikipedia and books …","url":["https://dspace.library.uu.nl/bitstream/handle/1874/406113/6400507_JorisAlbers_Thesis.pdf?sequence=1"]}
{"year":"2021","title":"Comparing Encoder-Decoder Architectures for Neural Machine Translation: A Challenge Set Approach","authors":["C Doan - 2021"],"snippet":"Machine translation (MT) as a field of research has known significant advances in recent years, with the increased interest for neural machine translation (NMT). By combining deep learning with translation, researchers have been able to deliver …","url":["https://ruor.uottawa.ca/bitstream/10393/42936/1/Doan_Coraline_2021_thesis.pdf"]}
{"year":"2021","title":"Comparing general and specialized word embeddings for biomedical named entity recognition","authors":["RE Ramos-Vargas, I Román-Godínez, S Torres-Ramos - PeerJ Computer Science, 2021"],"snippet":"… 01-14 Received 2020-11-05 Academic Editor Susan Gauch Subject Areas Bioinformatics, Artificial Intelligence, Computational Linguistics Keywords Word embeddings, BioNER, BiLSTM-CRF, DrugBank, MedLine, Pyysalo …","url":["https://peerj.com/articles/cs-384/"]}
{"year":"2021","title":"Comparing the Performance of NLP Toolkits and Evaluation measures in Legal Tech","authors":["MZ Khan, J Mitrovic, JMPDM Granitzer - 2021"],"snippet":"Page 1. Lehrstuhl für Data Science Comparing the Performance of NLP Toolkits and Evaluation measures in Legal Tech Masterarbeit von Muhammad Zohaib Khan Supervised By: Prof. Dr. Jelena Mitrovic 1. Prüfer 2. Prüfer …","url":["https://www.academia.edu/download/65887417/Deep_Neural_Language_Modelling_in_Law.pdf"]}
{"year":"2021","title":"Comparing Traditional and Neural Approaches for Detecting Health-Related Misinformation","authors":["D Elsweiler - … IR Meets Multilinguality, Multimodality, and Interaction …","M Fernández-Pichel, DE Losada, JC Pichel…"],"snippet":"… Table 1 reports the main statistics of the resulting datasets. We also tested classifiers for the task of distinguishing between useful documents for non-expert end users (ie, trustworthy and readable) and non-useful …","url":["http://persoal.citius.usc.es/jcpichel/docs/2021_CLEF_MFernandezPichel.pdf","https://books.google.de/books?hl=en&lr=lang_en&id=p9FCEAAAQBAJ&oi=fnd&pg=PA78&dq=commoncrawl&ots=eNycpv3vEv&sig=v7CAPFEmV26pL2Lhj2R2t581gZ0"]}
{"year":"2021","title":"Comparison of Czech Transformers on Text Classification Tasks","authors":["J Lehečka, J Švec - arXiv preprint arXiv:2107.10042, 2021"],"snippet":"… Researchers from Facebook have published multilingual XLM-RoBERTa model [3] pre-trained on one hundred languages (including Czech), using more than two terabytes of filtered Common Crawl data … 2. 1https …","url":["https://arxiv.org/pdf/2107.10042"]}
{"year":"2021","title":"Compilation and Validation of a Large Fake News Dataset in Hungarian","authors":["M Gencsi, Z Bodó, A Szenkovits - 2021 IEEE 19th International Symposium on …, 2021"],"snippet":"… The huBERT model was trained on the Hungarian subset of the Common Crawl and a snapshot of the Hungarian Wikipedia, while the multilingual model was trained on the top 104 languages with the largest Wikipedias, among them also …","url":["https://ieeexplore.ieee.org/abstract/document/9582484/"]}
{"year":"2021","title":"Comprehensive analysis of embeddings and pre-training in NLP","authors":["JK Tripathy, SC Sethuraman, MV Cruz, A Namburu… - Computer Science Review, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S1574013721000733"]}
{"year":"2021","title":"Comprehensive Evaluation of Word Embeddings for Highly Inflectional Language","authors":["P Drozda, K Sopyla, J Lewalski - International Conference on Computational …, 2021"],"snippet":"… The obtained results showed that in terms of accuracy the Facebook fasttext model learned on the Common Crawl collection should be considered the best model under assumptions of experimental session. Keywords. Word …","url":["https://link.springer.com/chapter/10.1007/978-3-030-88113-9_48"]}
{"year":"2021","title":"Comprehensive Multi-Modal Interactions for Referring Image Segmentation","authors":["K Jain, V Gandhi - arXiv preprint arXiv:2104.10412, 2021"],"snippet":"… 576. At 448 × 448 resolution, H = W = 14 and at 576 × 576 resolution, H = W = 18. We use GLoVe embeddings [17] pre-trained on Common Crawl 840B tokens to initialize word embedding for words in the expressions. The …","url":["https://arxiv.org/pdf/2104.10412"]}
{"year":"2021","title":"Compression, Transduction, and Creation: A Unified Framework for Evaluating Natural Language Generation","authors":["M Deng, B Tan, Z Liu, EP Xing, Z Hu - arXiv preprint arXiv:2109.06379, 2021"],"snippet":"Page 1. Compression, Transduction, and Creation: A Unified Framework for Evaluating Natural Language Generation Mingkai Deng1∗, Bowen Tan1∗, Zhengzhong Liu1,2, Eric P. Xing1,2,3, Zhiting Hu4 1Carnegie Mellon University …","url":["https://arxiv.org/pdf/2109.06379"]}
{"year":"2021","title":"Computational analysis and synthesis of song lyrics","authors":["P Březinová - 2021"],"snippet":"… It uses CMUdict for phonetic transcription, analyzes CommonCrawl8 web data repository for forced rhymes, Google Books Ngrams (Weiss [2015]) for building language model, and WordNet 3.0 (Pearson et al. [2005]) for semantic relations …","url":["https://dspace.cuni.cz/bitstream/handle/20.500.11956/147665/120397406.pdf?sequence=1"]}
{"year":"2021","title":"Computational Challenges for Artificial Intelligence and Machine Learning in Environmental Research","authors":["M Werner, G Dax, M Laass - INFORMATIK 2020, 2021"],"snippet":"… This includes news streams, social media messages, human-curated knowledge such as OpenStreetMap and Wikipedia, opinionated data sources such as blog posts from certain platforms, or blind web scale data collections such as common crawl …","url":["https://dl.gi.de/bitstream/handle/20.500.12116/34809/C21-1.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Computational filling of curatorial gaps in a fine arts exhibition","authors":["A Flexer"],"snippet":"… Please note that we translate all keywords from German to English for this paper. We use the German fasttext5 word em- bedding, which has been trained on about 3 million words from the Wikipediaand 19 million words …","url":["https://computationalcreativity.net/iccc21/wp-content/uploads/2021/09/ICCC_2021_paper_75reduced.pdf"]}
{"year":"2021","title":"Computational methods to understand the association between emojis and emotions","authors":["AAM Shoeb - 2021"],"snippet":"Page 1. © 2021 Abu Awal Md Shoeb ALL RIGHTS RESERVED Page 2. COMPUTATIONAL METHODS TO UNDERSTAND THE ASSOCIATION BETWEEN EMOJIS AND EMOTIONS By ABU AWAL MD SHOEB A dissertation submitted to the School of Graduate Studies …","url":["https://rucore.libraries.rutgers.edu/rutgers-lib/65975/PDF/1/"]}
{"year":"2021","title":"Computer Science Review","authors":["JK Tripathy, SC Sethuraman, MV Cruz, V Vijayakumar - 2021"],"snippet":"abstract The amount of data and computing power has drastically increased over the last decade, which leads to the development of several new fronts in the field of Natural Language Processing (NLP). In addition to that, the entanglement of …","url":["https://www.researchgate.net/profile/Mangalraj-Poobalasubramanian/publication/355132427_Comprehensive_analysis_of_embeddings_and_pre-training_in_NLP/links/6164f98e1eb5da761e836888/Comprehensive-analysis-of-embeddings-and-pre-training-in-NLP.pdf"]}
{"year":"2021","title":"CoMSum and SIBERT: A Dataset and Neural Model for Query-Based Multi-document Summarization","authors":["S Kulkarni, S Chammas, W Zhu, F Sha, E Ie - International Conference on Document …, 2021"],"snippet":"… We use the cleaned Common Crawl (CC) corpus [32] to source relevant web documents that are diverse and multi-faceted for generating Natural Questions (NQ) (long-form) answers [21]. Figure 1 illustrates the overall procedure …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86331-9_6"]}
{"year":"2021","title":"Concept-Based Label Embedding via Dynamic Routing for Hierarchical Text Classification","authors":["X Wang, L Zhao, B Liu, T Chen, F Zhang, D Wang"],"snippet":"… Hyper-parameters are tuned on a validation set by grid search. We take Stanford's publicly available GloVe 300-dimensional embeddings trained on 42 billion tokens from Common Crawl (Pennington et al., 2014) as initialization for word em- beddings …","url":["https://aclanthology.org/2021.acl-long.388.pdf"]}
{"year":"2021","title":"Confused by Path: Analysis of Path Confusion Based Attacks","authors":["SA Mirheidari - 2020"],"snippet":"… 93 iii Page 10. Page 11. List of Tables 4.1 Sample Grouped Web pages. . . . . 29 4.2 Narrowing down the Common Crawl to the candidate set used in our analysis (from left to right). . . . 36 4.3 Vulnerable pages and sites in the candidate set …","url":["https://iris.unitn.it/retrieve/handle/11572/280512/382175/phd_unitn_Seyed%20Ali_Mirheidari.pdf"]}
{"year":"2021","title":"ConRPG: Paraphrase Generation using Contexts as Regularizer","authors":["Y Meng, X Ao, Q He, X Sun, Q Han, F Wu, J Li - arXiv preprint arXiv:2109.00363, 2021"],"snippet":"… We implement the above models, ie p(−→ci|c<i,c>i), p(←−ci|c<i,c>i), p(c<i|ci,c>i), p(c>i|c<i,ci) based on the SEQ2SEQ structure on a subset of CommonCrawl containing 10 billion tokens in total. We use Transformers …","url":["https://arxiv.org/pdf/2109.00363"]}
{"year":"2021","title":"Consumer Health Search at CLEF eHealth 2021","authors":["L Goeuriot, G Pasi, H Suominen, E Bassani… - CLEF 2021 Evaluation Labs …, 2021"],"snippet":"… The task generated a new representative web corpus including web pages acquired from a 2021 CommonCrawl and social media content from Twitter and Reddit, along with a new collection of 55 manually generated layperson …","url":["http://ceur-ws.org/Vol-2936/paper-62.pdf"]}
{"year":"2021","title":"Contingency of semantic generalization on episodic specificity varies across development","authors":["CT Ngo, SL Benear, H Popal, IR Olson, NS Newcombe - Current Biology, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0960982221004619"]}
{"year":"2021","title":"Contrastive finetuning of generative language models for informal premise selection","authors":["JM Han, T Xu, S Polu, A Neelakantan, A Radford"],"snippet":"… contrastive training. • GPT-3 style pretraining. The model is pretrained for 300B tokens on the same data (a mix of filtered CommonCrawl, WebText, books, and Wikipedia) as GPT-3 [2]. • WebMath pretraining. Starting from the …","url":["http://aitp-conference.org/2021/abstract/paper_21.pdf"]}
{"year":"2021","title":"Control Prefixes for Text Generation","authors":["J Clive, K Cao, M Rei - arXiv preprint arXiv:2110.08329, 2021"],"snippet":"Prompt learning methods adapt pre-trained language models to downstream applications by using a task-specific prompt together with the input. Most of the current work on prompt learning in text generation relies on a shared dataset-level …","url":["https://arxiv.org/pdf/2110.08329"]}
{"year":"2021","title":"Conversations Powered by Cross-Lingual Knowledge","authors":["W Sun, C Meng, Q Meng, Z Ren, P Ren, Z Chen… - 2021"],"snippet":"Page 1. Conversations Powered by Cross-Lingual Knowledge Weiwei Sun1* Chuan Meng1* Qi Meng2 Zhaochun Ren1† Pengjie Ren1† Zhumin Chen1 Maarten de Rijke3,4 1Shandong University, Qingdao, China 2Microsoft …","url":["https://staff.fnwi.uva.nl/m.derijke/wp-content/papercite-data/pdf/sun-2021-conversations.pdf"]}
{"year":"2021","title":"CopyCat: Near-Duplicates within and between the ClueWeb and the Common Crawl","authors":["M Fröbe, J Bevendorff, L Gienapp, M Völske, B Stein… - 2021"],"snippet":"The amount of near-duplicates in web crawls like the ClueWeb or Common Crawl demands from their users either to develop a preprocessing pipeline for deduplication, which is costly both computationally and in person hours, or accepting …","url":["https://webis.de/downloads/publications/papers/froebe_2021.pdf"]}
{"year":"2021","title":"Corpulyzer: A Novel Framework for Building Low Resource Language Corpora","authors":["B Tahir, MA Mehmood - IEEE Access"],"snippet":"… Leveraging dataset from Common Crawl Corpus (CCC), first, we prepare a list of seed URLs by filtering the Urdu language webpages … INDEX TERMS Common crawl, web crawling, text corpus, corpus analysis, regional languages corpora …","url":["https://ieeexplore.ieee.org/iel7/6287639/9312710/09316706.pdf"]}
{"year":"2021","title":"Corpus and Baseline Model for Domain-Specific Entity Recognition in German","authors":["S Torge, W Hahn, R Jäkel, WE Nagel - 2020 6th IEEE Congress on Information …, 2021"],"snippet":"… Deepset AI4 FastText Wikipedia 100 Deepset AI GloVe Wikipedia 300 FastText [26] FastText Wikipedia, 300 Common Crawl Kyubyong5 Word2Vec Wikipedia 300 SB Tweets s[27] FastText 50 Mio Tweets 100 SB Tweets l[27] FastText 50 Mio Tweets 300 …","url":["https://ieeexplore.ieee.org/abstract/document/9357189/"]}
{"year":"2021","title":"COVIDSenti: A Large-Scale Benchmark Twitter Data Set for COVID-19 Sentiment Analysis","authors":["P BALAJI","U Naseem, I Razzak, M Khushi, PW Eklund, J Kim - IEEE Transactions on …, 2021"],"snippet":"… Similarly, for word embeddings, pretrained Word2Vec, GloVe, and fastText embeddings trained on Common Crawl and Wikipedia are used and have 300D vectors. In addition, we used hybrid models, such as hybrid ranking (HyRank) and …","url":["https://ieeexplore.ieee.org/abstract/document/9340540/","https://sist.sathyabama.ac.in/sist_naac/documents/1.3.4/1822-b.tech-it-batchno-359.pdf"]}
{"year":"2021","title":"CRASS: A Novel Data Set and Benchmark to Test Counterfactual Reasoning of Large Language Models","authors":["J Frohberg, F Binder - arXiv preprint arXiv:2112.11941, 2021"],"snippet":"We introduce the CRASS (counterfactual reasoning assessment) data set and benchmark utilizing questionized counterfactual conditionals as a novel and powerful tool to evaluate large language models. We present the data set design …","url":["https://arxiv.org/pdf/2112.11941"]}
{"year":"2021","title":"Creation, Enrichment and Application of Knowledge Graphs","authors":["S Gottschalk - 2021"],"snippet":"Page 1. CREATION, ENRICHMENT AND APPLICATION OF KNOWLEDGE GRAPHS Von der Fakultät für Elektrotechnik und Informatik der Gottfried Wilhelm Leibniz Universität Hannover zur Erlangung des Grades DOKTOR …","url":["https://www.repo.uni-hannover.de/bitstream/handle/123456789/11125/phd_thesis_gottschalk.pdf?sequence=1"]}
{"year":"2021","title":"CrisisBench: Benchmarking Crisis-related Social Media Datasets for Humanitarian Information Processing","authors":["F Alam, H Sajjad, M Imran, F Ofli - 2021"],"snippet":"… 2017). fastText: For the fastText (Joulin et al. 2017), we used pretrained embeddings trained on Common Crawl, which is re- leased by fastText for English. Transformer models: Pre-trained models have achieved …","url":["https://mimran.me/papers/CrisisBench_Benchmarking_Crisis_related_Social_Media_Datasets_ICWSM21.pdf"]}
{"year":"2021","title":"Cross-Domain Transfer of Generative Explanations Using Text-to-Text Models","authors":["KF Erliksson, A Arpteg, M Matskin, AH Payberah - International Conference on …, 2021"],"snippet":"… An analysis of BERT'S attention. In: ACL Blackbox NLP Workshop (2019)Google Scholar. 7. Common Crawl. https://www.commoncrawl.org. 8. Danilevsky, M., et al.: A survey of the state of explainable AI for natural language processing …","url":["https://link.springer.com/chapter/10.1007/978-3-030-80599-9_8"]}
{"year":"2021","title":"Cross-language Information Retrieval","authors":["P Galuščáková, DW Oard, S Nair - arXiv preprint arXiv:2111.05988, 2021"],"snippet":"Two key assumptions shape the usual view of ranked retrieval: (1) that the searcher can choose words for their query that might appear in the documents that they wish to see, and (2) that ranking retrieved documents will suffice because the searcher …","url":["https://arxiv.org/pdf/2111.05988"]}
{"year":"2021","title":"Cross-Language Learning for Entity Matching","authors":["R Peeters, C Bizer - arXiv preprint arXiv:2110.03338, 2021"],"snippet":"… We further use a German BERT model pre-trained on a di- verse set of German data ('bert-base-german-dbmdz-uncased'), including the German Wikipedia, parts of the Common Crawl, and EU Bookshop corpus, which …","url":["https://arxiv.org/pdf/2110.03338"]}
{"year":"2021","title":"Cross-lingual Capsule Network for Hate Speech Detection in Social Media","authors":["A Jiang, A Zubiaga - arXiv preprint arXiv:2108.03089, 2021"],"snippet":"… mBERT. Multilingual BERT6 (mBERT) is a variant of BERT [14] that was trained on 104 languages of Wikipedia; XLM-R. XLM-RoBERTa is a scaled cross-lingual sentence en- coder across 100 languages from Common Crawl [11]; …","url":["https://arxiv.org/pdf/2108.03089"]}
{"year":"2021","title":"Cross-lingual Fine-tuning for Abstractive Arabic Text Summarization","authors":["M Kahla, ZG Yang, A Novák - 2021"],"snippet":"… We only kept ar- ticles from the “Main/Top Stories” section, and filtered out all articles where either the main article text or the lead was too short or missing and items where the text is shorter than 4 times the 2https://www …","url":["http://nlpg.itk.ppke.hu/sites/default/files/publications/RANLP2021_Cross_lingual_fine_tuning_for_Abstractive_Arabic_Text_Summarization.pdf"]}
{"year":"2021","title":"Cross-lingual Hate Speech Detection using Transformer Models","authors":["T Tiţa, A Zubiaga - arXiv preprint arXiv:2111.00981, 2021"],"snippet":"Hate speech detection within a cross-lingual setting represents a paramount area of interest for all medium and large-scale online platforms. Failing to properly address this issue on a global scale has already led over time to morally questionable real-life …","url":["https://arxiv.org/pdf/2111.00981"]}
{"year":"2021","title":"Cross-lingual Language Model Pretraining for Retrieval","authors":["P Yu, H Fei, P Li - Proceedings of the Web Conference 2021, 2021"],"snippet":"… (MLM) and next sentence prediction (NSP) objectives [13] on the top 102 languages with the largest Wikipedia dumps. (v) XLM-R. We use the public checkpoint9 [8]. It was originally pretrained with the MLM objective on the CommonCrawl corpus in 100 languages …","url":["https://dl.acm.org/doi/abs/10.1145/3442381.3449830"]}
{"year":"2021","title":"Cross-Lingual Text Classification of Transliterated Hindi and Malayalam","authors":["J Krishnan, A Anastasopoulos, H Purohit, H Rangwala - arXiv preprint arXiv …, 2021"],"snippet":"… The improvement produced by our model is due to the fact that the pre-trained models have likely not seen that many transliterations in these languages. mBERT is trained using data from Wikipedia, while XLM-R uses Common Crawl …","url":["https://arxiv.org/pdf/2108.13620"]}
{"year":"2021","title":"Cross-lingual Transfer for Text Classification with Dictionary-based Heterogeneous Graph","authors":["N Chairatanakul, N Sriwatanasakdi… - arXiv preprint arXiv …, 2021"],"snippet":"… FastText (Bojanowski et al., 2017): a taskindependent word embedding trained on Wikipedia and Common Crawl only of a target language. • RCSLS (Joulin et al., 2018): a CLWE trained on Wikipedia. We used public FastText4 version …","url":["https://arxiv.org/pdf/2109.04400"]}
{"year":"2021","title":"Cross-lingual transfer of abstractive summarizer to less-resource language","authors":["A Žagar, M Robnik-Šikonja - Journal of Intelligent Information Systems, 2021"],"snippet":"Automatic text summarization extracts important information from texts and presents the information in the form of a summary. Abstractive summarization app.","url":["https://link.springer.com/article/10.1007/s10844-021-00663-8"]}
{"year":"2021","title":"Cross-lingual Transferring of Pre-trained Contextualized Language Models","authors":["Z Li, K Parnow, H Zhao, Z Zhang, R Wang, M Utiyama… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. ARXIV, DECEMBER 2020 1 Cross-lingual Transferring of Pre-trained Contextualized Language Models Zuchao Li, Kevin Parnow, Hai Zhao, Zhuosheng Zhang, Rui Wang, Masao Utiyama, and Eiichiro Sumita Abstract …","url":["https://arxiv.org/pdf/2107.12627"]}
{"year":"2021","title":"Cross-Lingual Word Embedding Refinement by $\\ell_ {1} $ Norm Optimisation","authors":["X Peng, C Lin, M Stevenson - arXiv preprint arXiv:2104.04916, 2021"],"snippet":"… Tibetan family). News-Embs (Artetxe et al., 2018a): embeddings trained on a multilingual News text collection, ie, Page 5. the WaCKy Crawl of {EN, DE, IT}, the Common Crawl of FI, and the WMT News Crawl of ES. News-Embs …","url":["https://arxiv.org/pdf/2104.04916"]}
{"year":"2021","title":"Cross-Lingual Word Embedding Refinement by ℓ1 Norm Optimisation Norm Optimisation","authors":["X Peng, C Lin, M Stevenson - Proceedings of the 2021 Conference of the North …, 2021"],"snippet":"… family). News-Embs (Artetxe et al., 2018a): embeddings trained on a multilingual News text collection, ie, Page 5. 2694 the WaCKy Crawl of {EN, DE, IT}, the Common Crawl of FI, and the WMT News Crawl of ES. News-Embs …","url":["https://www.aclweb.org/anthology/2021.naacl-main.214.pdf"]}
{"year":"2021","title":"Cross-lingual word embedding refinement by ℓ1 norm optimisation","authors":["X Peng, C Lin, M Stevenson - arXiv, 2021"],"snippet":"… Tibetan family). News-Embs (Artetxe et al., 2018a): embeddings trained on a multilingual News text collection, ie, Page 6. the WaCKy Crawl of {EN, DE, IT}, the Common Crawl of FI, and the WMT News Crawl of ES. News-Embs …","url":["https://eprints.whiterose.ac.uk/173333/1/2104.04916v1.pdf"]}
{"year":"2021","title":"Cross-Modal Progressive Comprehension for Referring Segmentation","authors":["S Liu, T Hui, S Huang, Y Wei, B Li, G Li - IEEE Transactions on Pattern Analysis and …, 2021"],"snippet":"Page 1. 0162-8828 (c) 2021 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This …","url":["https://ieeexplore.ieee.org/abstract/document/9430750/"]}
{"year":"2021","title":"CTNet: Conversational Transformer Network for Emotion Recognition","authors":["Z Lian, B Liu, J Tao - IEEE/ACM Transactions on Audio, Speech, and …, 2021"],"snippet":"… utterances. Specifically, we convert transcripts into the concatenation of vectors of the constituent words. These vectors are the public available word vectors trained on the Common Crawl and Wikipedia datasets [44]. Totally …","url":["https://ieeexplore.ieee.org/abstract/document/9316758/"]}
{"year":"2021","title":"CUNI systems for WMT21: Terminology translation Shared Task","authors":["J Jon, M Novák, JP Aires, D Variš, O Bojar - arXiv preprint arXiv:2109.09350, 2021"],"snippet":"… 3.1 Data We used all English-French corpora allowed by the organizers, aside from Paracrawl (with the exception of one model, which is marked). Namely this means Europarl v10, Common Crawl, UN Parallel Corpus …","url":["https://arxiv.org/pdf/2109.09350"]}
{"year":"2021","title":"Current Advances in Neural Networks","authors":["V Gallego, DR Insua - 2021"],"snippet":"Abstract This article reviews current advances and developments in neural networks. This requires recalling some of the earlier work in the field. We emphasize Bayesian approaches and their benefits compared to more standard maximum likelihood …","url":["https://www.annualreviews.org/doi/abs/10.1146/annurev-statistics-040220-112019"]}
{"year":"2021","title":"Data Augmentation for Speech Recognition in Maltese: A Low-Resource Perspective","authors":["C Mena, A DeMarco, C Borg, L van der Plas, A Gatt - arXiv preprint arXiv:2111.07793, 2021"],"snippet":"Developing speech technologies is a challenge for low-resource languages for which both annotated and raw speech data is sparse. Maltese is one such language. Recent years have seen an increased interest in the computational processing of …","url":["https://arxiv.org/pdf/2111.07793"]}
{"year":"2021","title":"Data augmentation impact on domain-specific text summarization","authors":["A LAIFA, C CRUZ, L Gautier"],"snippet":"… Association for Computational Linguistics, Seattle / Virtual, United States (2020) 4. Lin C., Text Summarization Branches Out, 74–81. Association for Computational Linguistics, Barcelona, Spain (2004) 5. Common Crawl Foundation, https://commoncrawl.org/","url":["https://www.researchgate.net/profile/Abdelghani-Laifa/publication/351935366_Data_augmentation_impact_on_domain-specific_text_summarization/links/60b0b492458515bfb0ac029e/Data-augmentation-impact-on-domain-specific-text-summarization.pdf"]}
{"year":"2021","title":"Data Efficient Deep Learning Models for Text Classification","authors":["R Garreta"],"snippet":"Page 1. Universidad de la República PEDECIBA Informática Master Thesis Data Efficient Deep Learning Models for Text Classification Author: Raúl Garreta Supervisors …","url":["https://www.fing.edu.uy/inco/grupos/pln/publicaciones/tesis_garreta.pdf"]}
{"year":"2021","title":"Data Filtering using Cross-Lingual Word Embeddings","authors":["C Herold, J Rosendahl, J Vanvinckenroye, H Ney - … of the 2021 Conference of the …, 2021"],"snippet":"… The De→En training data consists of the corpora Commoncrawl, Europarl, Rapid and ParaCrawl from the WMT 2019 news translation task2. We use the czeng 1.7 corpus3 from the WMT 2018 news translation task for En→Cs …","url":["https://www.aclweb.org/anthology/2021.naacl-main.15.pdf"]}
{"year":"2021","title":"Data science with Vadalog: Knowledge Graphs with machine learning and reasoning in practice","authors":["L Bellomarini, RR Fayzrakhmanov, G Gottlob… - Future Generation …, 2021"],"snippet":"… Such resources originate from data streams collected from internal systems (eg, Enterprise Resource Planning, Workflow Management, and Supply Chain Management) and external resources (eg, news, social media feeds and the Common Crawl 1 ). …","url":["https://www.sciencedirect.com/science/article/pii/S0167739X21004179"]}
{"year":"2021","title":"Dataset for Identification of Homophobia and Transophobia in Multilingual YouTube Comments","authors":["BR Chakravarthi, R Priyadharshini, R Ponnusamy… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Abstract The increased proliferation of abusive content on social media platforms has a negative impact on online users. The dread, dislike, discomfort, or mistrust of lesbian, gay, transgender or bisexual persons is de- fined as homophobia/transphobia …","url":["https://arxiv.org/pdf/2109.00227"]}
{"year":"2021","title":"DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing","authors":["P He, J Gao, W Chen - arXiv preprint arXiv:2111.09543, 2021"],"snippet":"This paper presents a new pre-trained language model, DeBERTaV3, which improves the original DeBERTa model by replacing mask language modeling (MLM) with replaced token detection (RTD), a more sample-efficient pre-training task. Our …","url":["https://arxiv.org/pdf/2111.09543"]}
{"year":"2021","title":"Debiasing Multilingual Word Embeddings: A Case Study of Three Indian Languages","authors":["S Bansal, V Garimella, A Suhane, A Mukherjee - arXiv preprint arXiv:2107.10181, 2021"],"snippet":"… We do not plan to release non-anonymized bios-data since it is already available for download from the commoncrawl portal. Page 4 … One of the main challenges here is the absence of appropriate datasets (bios) in Indic languages on Common Crawl dumps …","url":["https://arxiv.org/pdf/2107.10181"]}
{"year":"2021","title":"Decentralised Privacy: A Distributed Ledger Approach","authors":["P Papadopoulos, N Pitropakis, WJ Buchanan - … , and Devices; Hussain, CM, Di Sia, P …, 2021"],"snippet":"Our world due to the technological progress became fast-paced and is constantly evolving, thus changing every single day. Consequently, the most valuable asset on earth is not gold or oil anymore but data. Big data companies try to take advantage …","url":["https://link.springer.com/content/pdf/10.1007/978-3-030-58675-1_58-1.pdf"]}
{"year":"2021","title":"Deciphering Speech: a Zero-Resource Approach to Cross-Lingual Transfer in ASR","authors":["O Klejch, E Wallington, P Bell - arXiv preprint arXiv:2111.06799, 2021"],"snippet":"… To prevent this issue we always used a language model trained on CommonCrawl data when deciphering and later re-decoding the training data. The CommonCrawl language models were trained on approximately 100M tokens and …","url":["https://arxiv.org/pdf/2111.06799"]}
{"year":"2021","title":"Deduplicating Training Data Makes Language Models Better","authors":["K Lee, D Ippolito, A Nystrom, C Zhang, D Eck… - arXiv preprint arXiv …, 2021"],"snippet":"… models trained on CommonCrawl in- clude GPT-3 (Brown et al., 2020) with the addition of book datasets, GROVER (Zellers et al., 2019) on a restricted subset filtered to news domains called RealNews, and T5 (Raffel et al., 2020) …","url":["https://arxiv.org/pdf/2107.06499"]}
{"year":"2021","title":"Deep artificial neural networks reveal a distributed cortical network encoding propositional sentence-level meaning","authors":["AJ Anderson, D Kiela, JR Binder, L Fernandino… - Journal of Neuroscience, 2021"],"snippet":"… represents individual words as 300-dimensional floating point vectors derived by factorizing a word co- 286 occurrence matrix (vocabulary size is 2.2 million words and co-occurrences were measured across 840 287 billion tokens …","url":["https://www.jneurosci.org/content/early/2021/03/22/JNEUROSCI.1152-20.2021.abstract"]}
{"year":"2021","title":"Deep contextualized text representation and learning for fake news detection","authors":["M Samadi, M Mousavian, S Momtazi - Information Processing & Management, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0306457321002077"]}
{"year":"2021","title":"Deep Learning and Computational Chemistry","authors":["T James, D Hristozov - Artificial Intelligence in Drug Design, 2022"],"snippet":"… For example, ImageNet contains approximately 1.2 million labeled images [99] and Common Crawl, which contributes the bulk of the training data for GPT-3, nearly one trillion words [100]. The initial training phase for AlphaGo likewise utilized a set …","url":["https://link.springer.com/protocol/10.1007/978-1-0716-1787-8_5"]}
{"year":"2021","title":"Deep Learning for Blocking in Entity Matching: A Design Space Exploration","authors":["H Li, N Tang, M Ouzzani, Y Govind, D Paulsen, G Fung…"],"snippet":"… Popular word-level embeddings such as word2vec [52] and GloVe [67] and characterlevel embeddings such as fastText [9] often come with pre-trained embeddings that are trained on a large generic external corpus such as Wikipedia, Common Crawl or PubMed …","url":["http://da.qcri.org/ntang/pubs/deepblocker.pdf"]}
{"year":"2021","title":"Deep learning for sentence clustering in essay grading support","authors":["LH Chang, I Rastas, S Pyysalo, F Ginter - arXiv preprint arXiv:2104.11556, 2021"],"snippet":"… [16] and refer readers to this paper for further details of the embeddings. These embedding were induced using the implementation of the skip-gram algorithm [22] in the word2vec software package on Finnish Common Crawl data …","url":["https://arxiv.org/pdf/2104.11556"]}
{"year":"2021","title":"Deep Learning Transformer Architecture for Named Entity Recognition on Low Resourced Languages: State of the art results","authors":["R Hanslo - arXiv preprint arXiv:2111.00830, 2021"],"snippet":"… model trained on 100 languages uses 2.5 TB of CommonCrawl (CC) data [2]. From the 100 languages used by the XLM-R multilingual masked language model, it is noted that Afrikaans (af) and isiXhosa (xh) are included in the pre-training. …","url":["https://arxiv.org/pdf/2111.00830"]}
{"year":"2021","title":"Deep Learning With Anaphora Resolution for the Detection of Tweeters With Depression: Algorithm Development and Validation Study","authors":["A Wongkoblap, MA Vadillo, V Curcin - JMIR Mental Health, 2021"],"snippet":"Background: Mental health problems are widely recognized as a major public health challenge worldwide. This concern highlights the need to develop effective tools for detecting mental health disorders in the population. Social …","url":["https://mental.jmir.org/2021/8/e19824/"]}
{"year":"2021","title":"Deep Learning-based Sentiment Analysis of Facebook Data: The Case of Turkish Users","authors":["Ö Çoban, SA Özel, A İnan - The Computer Journal, 2021"],"snippet":"Abstract. Sentiment analysis (SA) is an essential task for many domains where it is crucial to know users' public opinion about events, products, brands, politi.","url":["https://academic.oup.com/comjnl/advance-article-abstract/doi/10.1093/comjnl/bxaa172/6095851"]}
{"year":"2021","title":"DeepEva: A Deep Neural Network Architecture for Assessing Sentence Complexity in Italian and English Languages","authors":["GL Bosco, G Pilato, D Schicchi - Array, 2021","GL Boscoa, G Pilato, D Schicchic"],"snippet":"Abstract Automatic Text Complexity Evaluation (ATE) is a research field that aims at creating new methodologies to make autonomous the process of the text complexity evaluation, that is the study of the text-linguistic features (eg, lexical, syntactical …","url":["https://iris.unipa.it/retrieve/handle/10447/524419/1257448/1-s2.0-S2590005621000424-main.pdf","https://www.sciencedirect.com/science/article/pii/S2590005621000424"]}
{"year":"2021","title":"Delving into Deep Imbalanced Regression","authors":["Y Yang, K Zha, YC Chen, H Wang, D Katabi - arXiv preprint arXiv:2102.09554, 2021"],"snippet":"Page 1. Delving into Deep Imbalanced Regression Yuzhe Yang 1 Kaiwen Zha 1 Ying-Cong Chen 1 Hao Wang 2 Dina Katabi 1 Abstract Real-world data often exhibit imbalanced distributions, where certain target values …","url":["https://arxiv.org/pdf/2102.09554"]}
{"year":"2021","title":"Democratic Backsliding and Media Responses to Government Repression: Machine Learning Evidence from Tanzania","authors":["FS Adiguzel, D Romero, E Wibbels"],"snippet":"… Using the list of international, regional and national domains, we first check GDELT and the Internet Archive for available links, pull the available web pages from the Common Crawl and from the websites directly. We then initialize Scrapy …","url":["https://mlp.trinity.duke.edu/assets/Tanzania_ML4P.pdf"]}
{"year":"2021","title":"Dense Events Grounding in Video","authors":["P Bao, Q Zheng, Y Mu - 2021"],"snippet":"… 2015) as previous methods to extract C3D video features on both datasets. And we use Glove (Jeffrey Pennington and Manning 2014) word embeddings pretrained on Common Crawl to represent each word in the sentences …","url":["http://www.muyadong.com/paper/3254_PeijunB.pdf"]}
{"year":"2021","title":"Depression Detection of Twitter Posters using Deep Learning with Anaphora Resolution","authors":["A Wongkoblap, MNR District"],"snippet":"… occurrent words as vectors [26]. GloVe was trained on several textual datasets eg, Wikipedia and common crawl (a copy of web content), and supported 50, 100, 200, and 300 dimensional vectors. Anaphora resolution is another …","url":["https://kclpure.kcl.ac.uk/portal/files/150024721/Journal_V2_270121_submitted.pdf"]}
{"year":"2021","title":"DESCGEN: A Distantly Supervised Datasetfor Generating Abstractive Entity Descriptions","authors":["W Shi, M Joshi, L Zettlemoyer - arXiv preprint arXiv:2106.05365, 2021"],"snippet":"… We further parse the HTML texts of the web pages and extract contents as source documents. Real News To expand the collection of source documents, we extract entity mentions in RealNews (Zellers et al., 2019) …","url":["https://arxiv.org/pdf/2106.05365"]}
{"year":"2021","title":"Design, analysis and implementation of advanced methodologies to measure the socio-economic impact of personal data in large online services","authors":["J González Cabañas - 2021"],"snippet":"The web ecosystem is enormous, and overall it is sustained by an intangible attribute that mainly supports the majority of free services: the exploitation of personal information. Over the years, concerns on how services use personal data …","url":["https://e-archivo.uc3m.es/bitstream/handle/10016/33628/tesis_jose_gonzalez_caba%C3%B1as_2021.pdf?sequence=1"]}
{"year":"2021","title":"Detecting Anatomical and Functional Connectivity Relations in Biomedical Literature via Language Representation Models","authors":["IB Ozyurt, J Menke, A Bandrowski, M Martone - … of the Second Workshop on Scholarly …, 2021"],"snippet":"… For word embeddings we have used 300 dimensional Common Crawl (840B tokens) trained GloVe (Pennington et al., 2014) vectors. The de- pendency parse trees for the input sentences were generated via Stanford CoreNLP (Manning et al., 2014) package …","url":["https://www.aclweb.org/anthology/2021.sdp-1.4.pdf"]}
{"year":"2021","title":"Detecting Discrete Emotions in Text Using Neural Networks","authors":["SA Seyeditabari - 2020"],"snippet":"Page 1. DETECTING DISCRETE EMOTIONS IN TEXT USING NEURAL NETWORKS by Seyed Armin Seyeditabari A dissertation submitted to the faculty of The University of North Carolina at Charlotte in partial fulfillment of the …","url":["https://search.proquest.com/openview/474e6c4af557600c03e4c198c711cf47/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Detecting ethnicity-targeted hate speech in Russian social media texts","authors":["E Pronoza, P Panicheva, O Koltsova, P Rosso - Information Processing & …, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0306457321001606"]}
{"year":"2021","title":"Detecting formal thought disorder by deep contextualized word representations","authors":["J Sarzynska-Wawer, A Wawer, A Pawlak… - Psychiatry Research, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0165178121004315"]}
{"year":"2021","title":"Detecting Hate Speech on Social Media Using Deep Learning Techniques","authors":["DO Idris, EO Akinola, S Olalekan"],"snippet":"… A FastText pre-trained model was obtained from Facebook's FastText website which contained 2 million word vectors with 300 dimensions trained with subword information on Common Crawl (consisting of a total of 600 billion tokens) …","url":["http://journals.ui.edu.ng/index.php/uijslictr/article/download/86/68"]}
{"year":"2021","title":"Detecting Online Risks and Supportive Interaction in Instant Messenger Conversations using Czech Transformers","authors":["O Sotolář, J Plhák, M Tkaczyk, M Lebedíková… - RASLAN 2021 Recent …, 2021"],"snippet":"We present a comparison of state-of-the-art models for text classification of Online Risks and Supportive Interaction in anonymized Instant Messenger conversations held in Czech. We compare the transformer models Czert, RobeCzech, and FERNET-C5 …","url":["https://nlp.fi.muni.cz/raslan/raslan21.pdf#page=27"]}
{"year":"2021","title":"Detecting Phishing Sites--An Overview","authors":["P Kalaharsha, BM Mehtre - arXiv preprint arXiv:2103.12739, 2021"],"snippet":"… Some of the common sources of datasets are shown below in Table 1. Alexa and Common crawl contains names of the legitimate sites which are likely to be used for phishing [62][63] … 2 Common-Crawl [63] 940 million URLs 2.8 billion webpages …","url":["https://arxiv.org/pdf/2103.12739"]}
{"year":"2021","title":"Detecting the target of sarcasm is hard: Really??","authors":["P Parameswaran, A Trotman, V Liesaputra, D Eyers - Information Processing & …, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0306457321000972"]}
{"year":"2021","title":"Detection and Morphological Analysis of Novel Russian Loanwords","authors":["Y Spektor - 2021"],"snippet":"… is compiled. The first dataset is the CC-100 Russian dataset (Conneau et al., 2020, Wenzek et al., 2020), which includes Common Crawl2 data from January 2018 to February 2019. A second … 3. 2 https://commoncrawl.org/ Page 16. 6 …","url":["https://academicworks.cuny.edu/cgi/viewcontent.cgi?article=5665&context=gc_etds"]}
{"year":"2021","title":"Detection of phishing attacks using probabilistic neural network with a novel training algorithm for reduced Gaussian kernels and optimal smoothing parameter …","authors":["S Priya, S Selvakumar - International Journal of Ad Hoc and Ubiquitous …, 2021"],"snippet":"Page 1. Int. J. Ad Hoc and Ubiquitous Computing, Vol. 36, No. 2, 2021 67 Detection of phishing attacks using probabilistic neural network with a novel training algorithm for reduced Gaussian kernels and optimal …","url":["https://www.inderscienceonline.com/doi/abs/10.1504/IJAHUC.2021.113381"]}
{"year":"2021","title":"Detection of Phishing Websites from URLs by using Classification Techniques on WEKA","authors":["B Geyik, K Erensoy, E Kocyigit - 2021 6th International Conference on Inventive …, 2021"],"snippet":"… We get the dataset from [21]. In this dataset, legitimate sites are collected from common-crawl and Alexa while phishing sites are collected from Phishtank. After some pre-processing steps 122,055 values 85,220 with …","url":["https://ieeexplore.ieee.org/abstract/document/9358642/"]}
{"year":"2021","title":"Development and Evaluation of Novel Ophthalmology Domain-Specific Neural Word Embeddings to Predict Visual Prognosis","authors":["S Wang, B Tseng, T Hernandez-Boussard - International Journal of Medical …, 2021"],"snippet":"… entity recognition, sentence classification, relation extraction, and other general natural language processing tasks,[8] and general domain GloVe (Global Vectors for Word Representation), pre-trained on large corpora of general …","url":["https://www.sciencedirect.com/science/article/pii/S1386505621000903"]}
{"year":"2021","title":"Development of a language model and opinion extraction for text analysis of online platforms","authors":["A Qudar, M Md - 2021"],"snippet":"Page 1. Lakehead University Knowledge Commons,http://knowledgecommons. lakeheadu.ca Electronic Theses and Dissertations Electronic Theses and Dissertations from 2009 2021 Development of a language model and opinion …","url":["https://knowledgecommons.lakeheadu.ca/bitstream/handle/2453/4822/QudarM2021m-1a.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"DFKI SLT at GermEval 2021: Multilingual Pre-training and Data Augmentation for the Classification of Toxicity in Social Media Comments","authors":["R Calizzano, M Ostendorff, G Rehm"],"snippet":"… It was trained on the Common Crawl corpus in 100 languages using masked language modeling. We choose XLM-RoBERTa instead of Multilingual BERT5 because XLM-RoBERTa outperforms Multilingual BERT on a variety of cross-lingual benchmarks …","url":["https://netlibrary.aau.at/obvukloa/periodical/titleinfo/6435231/full.pdf"]}
{"year":"2021","title":"Dictionary-based Debiasing of Pre-trained Word Embeddings","authors":["M Kaneko, D Bollegala - arXiv preprint arXiv:2101.09525, 2021"],"snippet":"… 3M words learned from Google News corpus (Mikolov et al., 2013a)), GloVe4 (300dimensional embeddings for ca. 2.1M words learned from the Common Crawl (Pennington et al., 2014)), and fastText5 (300-dimensional embeddings for ca …","url":["https://arxiv.org/pdf/2101.09525"]}
{"year":"2021","title":"Different Keystrokes for Different Folks: Visualizing Crowdworker Querying Behavior","authors":["R Benham, J Mackenzie, JS Culpepper, A Moffat - 2021"],"snippet":"… In this paper, we present the Common Crawl News Query Ex- plorer, 1 which enables information retrieval practitioners to ob- serve and compare human query formulation a posteriori … Common Crawl News Query Explorer …","url":["https://jmmackenzie.io/pdf/bm+21-chiir.pdf"]}
{"year":"2021","title":"Difficulty-as-improvement: The courage to keep going in the face of life's difficulties","authors":["V Yan, D Oyserman, G Kiper, M Atari - 2021"],"snippet":"… Our computational linguistic analyses of the “Common Crawl” corpus suggest that when people talk about difficulty they also talk about importance, impossibility, and improvement (Study 1). Studies 2-14 (total N = 2,378) use our brief difficulty-as-improvement …","url":["https://psyarxiv.com/f7epj/download?format=pdf"]}
{"year":"2021","title":"Digital humanities and web archives: Possible new paths for combining datasets","authors":["N Brügger - International Journal of Digital Humanities, 2021"],"snippet":"… Only a limited number of web archives are open to the wider public, most notably the US based Internet Archive, but also the Portuguese, the Croatian, the Australian, and the Icelandic web archives, and open collections such …","url":["https://link.springer.com/article/10.1007/s42803-021-00038-z"]}
{"year":"2021","title":"Discovering prerequisite relations from educational documents through word embeddings","authors":["F Gasparetti - Future Generation Computer Systems, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0167739X21003290"]}
{"year":"2021","title":"Discovering Representation Sprachbund For Multilingual Pre-Training","authors":["Y Fan, Y Liang, A Muzio, H Hassan, H Li, M Zhou… - arXiv preprint arXiv …, 2021"],"snippet":"… We use Wikipedia1 corpora (100 languages are included, total size 101GB) and a clean version of Common Crawl (CC)2 (89 languages are included, total size 2500GB) following Liang et al. (2020). 108 languages are included in our multilingual corpora …","url":["https://arxiv.org/pdf/2109.00271"]}
{"year":"2021","title":"Discovering X Degrees of Keyword Separation in a Fine Arts Collection","authors":["A Flexer"],"snippet":"… Please note that we translate all keywords from German to English for this paper. We use the German fasttext word embedding, which has been trained on about 3 million words from the Wikipediaand 19 million words from …","url":["https://www.researchgate.net/profile/Arthur-Flexer/publication/350721506_Discovering_X_Degrees_of_Keyword_Separation_in_a_Fine_Arts_Collection/links/606e9f15a6fdcc5f778cad52/Discovering-X-Degrees-of-Keyword-Separation-in-a-Fine-Arts-Collection.pdf"]}
{"year":"2021","title":"Discrete and Soft Prompting for Multilingual Models","authors":["M Zhao, H Schütze - arXiv preprint arXiv:2109.03630, 2021"],"snippet":"… Setup. We conduct all experiments using the pretrained XLM-RoBERTa-base model (Conneau et al., 2020) containing 270M parameters trained on 2.5 TB CommonCrawl data in 100 languages. We use PyTorch (Paszke et al., 2019) and the Hugprompt …","url":["https://arxiv.org/pdf/2109.03630"]}
{"year":"2021","title":"Discrete Cosine Transform as Universal Sentence Encoder","authors":["N Almarwani, M Diab - arXiv preprint arXiv:2106.00934, 2021"],"snippet":"… Specifically, we used a sample of 1 million parallel sentences from WMT'13 common-crawl data; this subset is the same one used in (Aldarmaki and Diab, 2019).4 To assess efficacy of the DCT models for the cross …","url":["https://arxiv.org/pdf/2106.00934"]}
{"year":"2021","title":"Discriminative Reranking for Neural Machine Translation","authors":["A Lee, M Auli, MA Ranzato"],"snippet":"… We follow the steps in Ng et al. (2019) for data preprocessing, including sentence deduplication, language identification filtering on all bitext and monolingual data (Joulin et al., 2017) and in- domain filtering (Moore …","url":["https://aclanthology.org/2021.acl-long.563.pdf"]}
{"year":"2021","title":"Dissociating Semantic and Phonemic Search Strategies in the Phonemic Verbal Fluency Task in early Dementia","authors":["H Lindsay, P Mueller, N Linz, R Zeghari, MM Mina… - Proceedings of the Seventh …, 2021"],"snippet":"… To obtain semantic word embeddings, the pretrained French fastText model is used. This model is trained on Common Crawl and Wikipedia corpora using the continuous bag of words (CBOW) algorithm with a negative sampling loss function …","url":["https://www.aclweb.org/anthology/2021.clpsych-1.4.pdf"]}
{"year":"2021","title":"DistilBERT-based Argumentation Retrieval for Answering Comparative Questions","authors":["A Alhamzeh, M Bouhaouel, E Egyed-Zsigmond… - Working Notes of CLEF, 2021"],"snippet":"… doi:10.1007/ 978-3-030-22948-1\\_5. [16] J. Bevendorff, B. Stein, M. Hagen, M. Potthast, Elastic ChatNoir: Search Engine for the ClueWeb and the Common Crawl, in: L. Azzopardi, A. Hanbury, G. Pasi, B. Piwowarski (Eds.), Advances in Information Retrieval …","url":["http://ceur-ws.org/Vol-2936/paper-209.pdf"]}
{"year":"2021","title":"DISTRIBUTIONAL INVESTIGATION OF SOME FREQUENT TURKISH DERIVATIONAL AFFIXES FOR EXPLORING THEIR SEMANTICS","authors":["GN Özdemir - 2021"],"snippet":"Page 1. DISTRIBUTIONAL INVESTIGATION OF SOME FREQUENT TURKISH DERIVATIONAL AFFIXES FOR EXPLORING THEIR SEMANTICS A THESIS SUBMITTED TO THE GRADUATE SCHOOL OF INFORMATICS OF THE …","url":["https://open.metu.edu.tr/bitstream/handle/11511/91631/gizemnur_ozdemir_tez.pdf"]}
{"year":"2021","title":"Diverse Distributions of Self-Supervised Tasks for Meta-Learning in NLP","authors":["T Bansal, K Gunasekaran, T Wang, T Munkhdalai… - arXiv preprint arXiv …, 2021"],"snippet":"… This performance drops even further (≈ 10%; r2c4) when the tasks are sampled from a different domain (common crawl) than the training domain … Figure 2 further shows that, for the Cluster tasks, constructing tasks from more diverse domain such …","url":["https://arxiv.org/pdf/2111.01322"]}
{"year":"2021","title":"DLJUST at SemEval-2021 Task 7: Hahackathon: Linking Humor and Offense","authors":["H Al-Omari, R Duwairi"],"snippet":"… Training the model on larger data set significantly improves the model performance using BookCurpus, English Wikipedia, CC-news with 63 million English news articles, OpenWebText, and Stories; which is a subset of CommonCrawl data …","url":["https://aclanthology.org/2021.semeval-1.155.pdf"]}
{"year":"2021","title":"Do Images really do the Talking? Analysing the significance of Images in Tamil Troll meme classification","authors":["SU Hegde, A Hande, R Priyadharshini, S Thavareesan… - arXiv preprint arXiv …, 2021"],"snippet":"… To address data imbalance during pretraining, the low-resource data are up- sampled while the high-resourceful data are downsampled to smoothen the data. The model is pretrained from scratch using the Wikipedia2, Common …","url":["https://arxiv.org/pdf/2108.03886"]}
{"year":"2021","title":"Do Syntactic Probes Probe Syntax? Experiments with Jabberwocky Probing","authors":["RH Maudslay, R Cotterell - Proceedings of the 2021 Conference of the North …, 2021"],"snippet":"Page 1. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 124–131 June 6–11, 2021. ©2021 Association for Computational Linguistics 124 …","url":["https://www.aclweb.org/anthology/2021.naacl-main.11.pdf"]}
{"year":"2021","title":"DOCmT5: Document-Level Pretraining of Multilingual Language Models","authors":["CH Lee, A Siddhant, V Ratnakar, M Johnson - arXiv preprint arXiv:2112.08709, 2021"],"snippet":"In this paper, we introduce DOCmT5, a multilingual sequence-to-sequence language model pre-trained with large scale parallel documents. While previous approaches have focused on leveraging sentence-level parallel data, we try to build …","url":["https://arxiv.org/pdf/2112.08709"]}
{"year":"2021","title":"Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus","authors":["J Dodge, M Sap, A Marasovic, W Agnew, G Ilharco…"],"snippet":"… In addition to C4.EN, we host the “uncleaned” version (C4.EN.NOCLEAN), which is the snapshot of Common Crawl identified as … 5https://github.com/ allenai/c4- documentation/discussions 6https://commoncrawl.org/, where …","url":["https://homes.cs.washington.edu/~msap/pdfs/dodge2021documentingC4.pdf"]}
{"year":"2021","title":"Documenting the English Colossal Clean Crawled Corpus","authors":["J Dodge, M Sap, A Marasovic, W Agnew, G Ilharco… - arXiv preprint arXiv …, 2021"],"snippet":"… been established. In this work we provide the first documentation for the Colossal Clean Crawled Corpus (C4; Raffel et al., 2020), a dataset created by applying a set of filters to a single snapshot of Common Crawl. We begin …","url":["https://arxiv.org/pdf/2104.08758"]}
{"year":"2021","title":"Does Robustness Improve Fairness? Approaching Fairness with Word Substitution Robustness Methods for Text Classification","authors":["Y Pruksachatkun, S Krishna, J Dhamala, R Gupta… - arXiv preprint arXiv …, 2021"],"snippet":"… We take 97,320 examples from the public leaderboard as the test set. • Bias in Bios (De-Arteaga et al., 2019) 3 is a dataset for occupation classification derived from Common Crawl corpus. It consists of 178,619 train and 91,917 test examples …","url":["https://arxiv.org/pdf/2106.10826"]}
{"year":"2021","title":"Does semantics aid syntax? An empirical study on named entity recognition and classification","authors":["X Zhong, E Cambria, A Hussain - Neural Computing and Applications, 2021"],"snippet":"Many researchers jointly model multiple linguistic tasks (eg, joint modeling of named entity recognition and named entity classification and joint modeli.","url":["https://link.springer.com/article/10.1007/s00521-021-05949-0"]}
{"year":"2021","title":"Domain-specific Taxonomy Enrichment based on Meta-Embeddings","authors":["M Tikhomirov, NV Loukachevitch - 2021"],"snippet":"In this paper we study the use of meta-embeddings approaches, which combine several source embeddings, for the taxonomy class prediction of new terms. We test the proposed approach in the informationsecurity domain in the task of enriching the …","url":["http://ceur-ws.org/Vol-3036/paper23.pdf"]}
{"year":"2021","title":"Don't Stop at the Top: Using Certificate Transparency Logs to Extend Domain Lists for Web Security Studies","authors":["F Marquardt, C Schmidt - 2020 IEEE 45th Conference on Local Computer …, 2020"],"snippet":"… They propose a system that collects massive amounts of records from various data sources, including data from top lists, blacklists, the Common Crawl project and DNS zone files. From these data sources, they were able to extract 199,110,841 unique domain names …","url":["https://ieeexplore.ieee.org/abstract/document/9314793/"]}
{"year":"2021","title":"DRo: A data-scarce mechanism to revolutionize the performance of DL-based Security Systems","authors":["M Sewak, SK Sahay, H Rathore - 2021 IEEE 46th Conference on Local Computer …, 2021"],"snippet":"… For pre-training a Transformer based TL language models like the OpenAI's GPT-3 [7] or the Google's BERT [5], billions to even trillions of (web) common-crawl documents could be easily and inexpensively made available …","url":["https://ieeexplore.ieee.org/abstract/document/9524929/"]}
{"year":"2021","title":"DS-TOD: Efficient Domain Specialization for Task Oriented Dialog","authors":["CC Hung, A Lauscher, SP Ponzetto, G Glavaš - arXiv preprint arXiv:2110.08395, 2021"],"snippet":"Recent work has shown that self-supervised dialog-specific pretraining on large conversational datasets yields substantial gains over traditional language modeling (LM) pretraining in downstream task-oriented dialog (TOD). These approaches, however …","url":["https://arxiv.org/pdf/2110.08395"]}
{"year":"2021","title":"DSGPT: Domain-Specific Generative Pre-Training of Transformers for Text Generation in E-commerce Title and Review Summarization","authors":["X Zhang, Y Jiang, Y Shang, Z Cheng, C Zhang, X Fan… - Proceedings of the 44th …, 2021"],"snippet":"… powerful. Pre-training a language model from a general corpus such as Wikipedia or the Common Crawl requires tremendous time and resource commitment, and can be wasteful if the downstream tasks are limited in variety …","url":["https://dl.acm.org/doi/pdf/10.1145/3404835.3463037"]}
{"year":"2021","title":"Dual-Objective Fine-Tuning of BERT for Entity Matching","authors":["R Peeters, C Bizer"],"snippet":"… WDC LSPC datasets: We use the training and test sets from the WDC Product Data Corpus for Large-scale Product Matching (WDC LSPC)1 [27] for the evaluation. The datasets were built by ex- tracting product offers from the Common Crawl …","url":["https://www.uni-mannheim.de/media/Einrichtungen/dws/Files_Research/Web-based_Systems/pub/Peeters-Bizer-Dual-Objective-Fine-Tuning-VLDB2021-preprint.pdf"]}
{"year":"2021","title":"Dual-State Capsule Networks for Text Classification","authors":["P Demotte, S Ranathunga - arXiv preprint arXiv:2109.04762, 2021"],"snippet":"Page 1. Dual-State Capsule Networks for Text Classification Piyumal Demotte, Surangika Ranathunga Department of Computer Science and Engineering, University of Moratuwa Katubedda 10400, Sri Lanka {piyumalanthony.16,surangika}@cse.mrt.ac.lk Abstract …","url":["https://arxiv.org/pdf/2109.04762"]}
{"year":"2021","title":"DuRecDial 2.0: A Bilingual Parallel Corpus for Conversational Recommendation","authors":["Z Liu, H Wang, ZY Niu, H Wu, W Che - arXiv preprint arXiv:2109.08877, 2021"],"snippet":"Page 1. DuRecDial 2.0: A Bilingual Parallel Corpus for Conversational Recommendation Zeming Liu1∗, Haifeng Wang2, Zheng-Yu Niu2, Hua Wu2, Wanxiang Che1† 1Research Center for Social Computing and Information …","url":["https://arxiv.org/pdf/2109.08877"]}
{"year":"2021","title":"Dynamic Knowledge Capitalization in Big Data Context: Monte Carlo Tree Search Based Reinforcement Learning","authors":["B Hirchoua, B Ouhbi, B Frikh - The International Conference on Artificial Intelligence …, 2021"],"snippet":"… 1. Fig. 1. Sub-trees examples alongside their local rewards. In order to demonstrate the effectiveness and efficiency of the proposed approach, a Common Crawl big data set is used for evaluation. Texts in this collection …","url":["https://link.springer.com/chapter/10.1007/978-3-030-76346-6_64"]}
{"year":"2021","title":"Dynamic Tuning and Weighting of Meta-learning for NMT Domain Adaptation","authors":["Z Song, Z Ma, K Qi, G Liu - International Conference on Artificial Neural Networks, 2021"],"snippet":"… 4 Experiments. 4.1 Datasets. Experiments are conducted on English-German parallel datasets. We use 6.4M sentence pairs to get a pre-trained model. The training data are sampled from CommonCrawl, Europarl, NewsCommentary, Wikititles and Rapid2019 …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86383-8_46"]}
{"year":"2021","title":"Easy-to-use combination of POS and BERT model for domain-specific and misspelled terms","authors":["A Benamar, M Bothua, C Grouin, A Vilnat"],"snippet":"… – Oscar [19] is a set of monolingual corpora extracted from Common Crawl. It was selected using a classification model for each language following the approach of [6] based on FastText [12]. The classifier was previously pretrained on Wikipedia …","url":["http://ceur-ws.org/Vol-3015/paper132.pdf"]}
{"year":"2021","title":"Echo chambers revisited: The (overwhelming) sharing of in-group politicians, pundits and media on Twitter","authors":["M Wojcieszak, A Casas, X Yu, J Nagler, JA Tucker - 2021"],"snippet":"Page 1. DRAFT Echo chambers revisited: The (overwhelming) sharing of in-group politicians, pundits and media on Twitter Magdalena Wojcieszak a,b , Andreu Casas c , Xudong Yu a , Jonathan Nagler d , and Joshua A. Tucker d …","url":["https://osf.io/xwc79/download?format=pdf"]}
{"year":"2021","title":"ECKPN: Explicit Class Knowledge Propagation Network for Transductive Few-shot Learning","authors":["C Chen, X Yang, C Xu, X Huang, Z Ma - Proceedings of the IEEE/CVF Conference on …, 2021"],"snippet":"… Specifically, we first leverage the GloVe (pretrained on a large text corpora with self-supervised constraint) [33] to obtain the d1-dimensional semantic embeddings of class labels. The Common Crawl version of the GloVe is …","url":["https://openaccess.thecvf.com/content/CVPR2021/papers/Chen_ECKPN_Explicit_Class_Knowledge_Propagation_Network_for_Transductive_Few-Shot_Learning_CVPR_2021_paper.pdf"]}
{"year":"2021","title":"EEGdenoiseNet: A benchmark dataset for end-to-end deep learning solutions of EEG denoising","authors":["H Zhang, M Zhao, C Wei, D Mantini, Z Li, Q Liu - Journal of Neural Engineering, 2021"],"snippet":"Page 1. Journal of Neural Engineering ACCEPTED MANUSCRIPT • OPEN ACCESS EEGdenoiseNet: A benchmark dataset for end-to-end deep learning solutions of EEG denoising To cite this article before publication: Haoming Zhang et al 2021 J. Neural Eng …","url":["https://iopscience.iop.org/article/10.1088/1741-2552/ac2bf8/pdf"]}
{"year":"2021","title":"Effective Seed-Guided Topic Labeling for Dataless Hierarchical Short Text Classification","authors":["Y Yang, H Wang, J Zhu, W Shi"],"snippet":"… We set the number of iterations to 8 as our model achieves competitive performance since then. For word embeddings, we employ the widely used GloVe Common Crawl6 [21], which contains 840B tokens, 2.2M vocab and 300d vectors …","url":["https://link.springer.com/content/pdf/10.1007/978-3-030-74296-6_21.pdf"]}
{"year":"2021","title":"Efficiency and Scalability of Large Search Architectures","authors":["M Siedlaczek - 2021"],"snippet":"Page 1. Page 2. 4/21/2021 04/25/2021 Page 3. iii Micro lm or other copies of this dissertation are obtainable from UMI Dissertation Publishing ProQuest CSA 789 E. Eisenhower Parkway PO Box 1346 Ann Arbor, MI 48106-1346 Page 4. iv V …","url":["https://search.proquest.com/openview/eac3a8ad7dcc3c4ed49f8beb9d39dfce/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Efficient Large Scale Language Modeling with Mixtures of Experts","authors":["M Artetxe, S Bhosale, N Goyal, T Mihaylov, M Ott… - arXiv preprint arXiv …, 2021"],"snippet":"Mixture of Experts layers (MoEs) enable efficient scaling of language models through conditional computation. This paper presents a detailed empirical study of how autoregressive MoE language models scale in comparison with dense models …","url":["https://arxiv.org/pdf/2112.10684"]}
{"year":"2021","title":"Efficient Passage Retrieval with Hashing for Open-domain Question Answering","authors":["I Yamada, A Asai, H Hajishirzi - arXiv preprint arXiv:2106.00882, 2021"],"snippet":"… the index of a common knowledge source (eg, Wikipedia) requires dozens of gigabytes.1 A reduction in the index size is critical for real-world QA that requires large knowledge sources such as scientific databases (eg, PubMed) …","url":["https://arxiv.org/pdf/2106.00882"]}
{"year":"2021","title":"Emerging trends: A gentle introduction to fine-tuning","authors":["KW Church, Z Chen, Y Ma - Natural Language Engineering, 2021"],"snippet":"The previous Emerging Trends article (Church et al., 2021. Natural Language Engineering 27(5), 631–645.) introduced deep nets to poets. Poets is an imperfect metaphor, intended as a gesture toward inclusion. The future for deep nets will …","url":["https://www.cambridge.org/core/services/aop-cambridge-core/content/view/C31D429D0928351D6A6692F8ECD1E7ED/S1351324921000322a.pdf/emerging_trends_a_gentle_introduction_to_finetuning.pdf"]}
{"year":"2021","title":"Emotionally informed hate speech detection: a multi-target perspective","authors":["P Chiril, EW Pamungkas, F Benamara, V Moriceau… - Cognitive Computation, 2021"],"snippet":"Hate Speech and harassment are widespread in online communication, due to users' freedom and anonymity and the lack of regulation provided by social me.","url":["https://link.springer.com/article/10.1007/s12559-021-09862-5"]}
{"year":"2021","title":"Empirical Best Practices On Using Product-Specific Schema. org","authors":["M Kejriwal, RK Selvam, CC Ni, N Torzec - Proceedings of the AAAI Conference on …, 2021"],"snippet":"… Keywords: E-commerce, Schema.org, Best Practices, Web Data Commons, Common Crawl, Markup, Big Data Abstract. Schema.org has experienced high growth in recent years. Structured descriptions of products embedded …","url":["https://ojs.aaai.org/index.php/AAAI/article/view/17816"]}
{"year":"2021","title":"Empirical Regularization for Synthetic Sentence Pairs in Unsupervised Neural Machine Translation","authors":["X Ai, B Fang - 2021"],"snippet":"Page 1. Empirical Regularization for Synthetic Sentence Pairs in Unsupervised Neural Machine Translation Xi Ai, Bin Fang Chongqing University barid.x.ai@gmail.com, fb@cqu.edu.cn Abstract UNMT tackles …","url":["https://www.aaai.org/AAAI21Papers/AAAI-520.AiXi.pdf"]}
{"year":"2021","title":"End-to-End Multimodal Clinical Depression Recognition using Deep Neural Networks: A comparative Analysis","authors":["M Muzammel, H Salam, A Othmani - Computer Methods and Programs in …, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0169260721005071"]}
{"year":"2021","title":"Energy-saving Cross-layer Optimization of Big Data Transfer Based on Historical Log Analysis","authors":["L Rodolph, MD Nine, L Di Tacchio, T Kosar - arXiv preprint arXiv:2104.01192, 2021"],"snippet":"… These include: (1) the small file size dataset consisting of 20,000 HTML files derived from the common crawl project [2]; (2) the medium file size dataset consisting of 5,000 image files derived from Flickr [4]; (3) the large …","url":["https://arxiv.org/pdf/2104.01192"]}
{"year":"2021","title":"Engineering Knowledge Graph from Patent Database","authors":["L Siddharth, L Blessing, KL Wood, J Luo - arXiv preprint arXiv:2106.06739, 2021"],"snippet":"… pdf Page 25. embedding models that were trained on Google News, Wikipedia, gigaword, and Common Crawl as the original data sources. The entries in Table 2 that mention these models are the resultant semantic networks …","url":["https://arxiv.org/pdf/2106.06739"]}
{"year":"2021","title":"Engineering Practical Lempel-Ziv Tries","authors":["D Arroyuelo, R Cánovas, J Fischer, D Köppl, M Löbel…"],"snippet":"Page 1. Engineering Practical Lempel-Ziv Tries ∗ Diego Arroyuelo1,2 Rodrigo Cánovas3 Johannes Fischer4 Dominik Köppl5 Marvin Löbel4 Gonzalo Navarro 1,6 Rajeev Raman7 1 Millennium Institute for Foundational Research on Data (IMFD), Chile …","url":["https://users.dcc.uchile.cl/~gnavarro/ps/jea21.pdf"]}
{"year":"2021","title":"ENHANCED REPRESENTATIONS AND EFFICIENT ANALYSIS OF SYNTACTIC DEPENDENCIES WITHIN AND BEYOND TREE STRUCTURES","authors":["T Shi - 2021"],"snippet":"Page 1. ENHANCED REPRESENTATIONS AND EFFICIENT ANALYSIS OF SYNTACTIC DEPENDENCIES WITHIN AND BEYOND TREE STRUCTURES A Dissertation Presented to the Faculty of the Graduate School of Cornell University …","url":["http://www.cs.cornell.edu/~tianze/papers/thesis.pdf"]}
{"year":"2021","title":"Enhanced sentiment extraction architecture for social media content analysis using capsule networks","authors":["P Demotte, K Wijegunarathna, D Meedeniya, I Perera - Multimedia Tools and …, 2021"],"snippet":"Recent research has produced efficient algorithms based on deep learning for text-based analytics. Such architectures could be readily applied to text-base.","url":["https://link.springer.com/article/10.1007/s11042-021-11471-1"]}
{"year":"2021","title":"Enhancing Context Through Contrast","authors":["K Ambilduke, A Shetye, D Bagade, R Bhagwatkar…"],"snippet":"Neural machine translation benefits from semantically rich representations. Considerable progress in learning such representations has been achieved by language modelling and mutual information maximization objectives using …","url":["https://preregister.science/papers_21neurips/18_paper.pdf"]}
{"year":"2021","title":"Enhancing Cross-lingual Semantic Annotations using Deep Network Sentence Embeddings","authors":["YC Lin, P Hoffmann, E Rahm - 2021"],"snippet":"… RoBERTa. XLM- R uses only MLM as training objective as RoBERTa (ie no TLM) but with multilingual corpus. It is trained on even larger dataset (the CommonCrawl corpus (Wenzek et al., 2019), 2.5 TB) and in 100 languages. XLM …","url":["https://dbs.uni-leipzig.de/file/HEALTHINF21_YCLIN.pdf"]}
{"year":"2021","title":"Enhancing Viewing Experience of Generated Visual Storylines for Promotional Videos","authors":["C Liu, H Yu, Z Shen, I Dixon, Y Yu, Z Gao, P Wang… - 2021 IEEE International …, 2021"],"snippet":"… ResNet50. Both models are trained on the DeepFashion dataset [21] until convergence2. Word embeddings are obtained from the pre-trained model on the 840B Common Crawl dataset. λi is set to 0.5. α, β, αc and βc are set to 0.4 …","url":["https://ieeexplore.ieee.org/abstract/document/9428292/"]}
{"year":"2021","title":"Ensembling of Distilled Models from Multi-task Teachers for Constrained Resource Language Pairs","authors":["A Hendy, EA Gad, M Abdelghaffar, JS ElMosalami… - arXiv preprint arXiv …, 2021"],"snippet":"… all the available monolingual data, eg NewsCrawl + CommonCrawl + Extended CommonCrawl for Hausa, and Extended CommonCrawl for both Xhosa and Zulu. For relatively … In addition to the NewsCrawl monolingual subset, we add a …","url":["https://arxiv.org/pdf/2111.13284"]}
{"year":"2021","title":"Enterprise Data Integration: On Extracting Data from HTML Tables","authors":["JC Roldán Salvador - 2020"],"snippet":"… layer is used. The authors learnt the network using 3 567 tables from 200 web sites, and evaluated the results on 60 678 ta- bles from 300 web sites; the documents were selected from the April 2016 Common Crawl. They also …","url":["https://idus.us.es/bitstream/handle/11441/105486/1/Rold%C3%A1n%20Salvador%2C%20Carlos%20tesis.pdf?sequence=1"]}
{"year":"2021","title":"ERNIE 3.0: Large-scale Knowledge Enhanced Pre-training for Language Understanding and Generation","authors":["Y Sun, S Wang, S Feng, S Ding, C Pang, J Shang, J Liu… - arXiv preprint arXiv …, 2021"],"snippet":"… However, these large-scale pre-trained language models with hundreds of billions of parameters are trained on plain texts. For example, the 175-billion-parameter GPT-3 is trained on a corpus with 570GB filtered texts from Common Crawl …","url":["https://arxiv.org/pdf/2107.02137"]}
{"year":"2021","title":"ERNIE-M: Enhanced Multilingual Representation by Aligning Cross-lingual Semantics with Monolingual Corpora","authors":["X Ouyang, S Wang, C Pang, Y Sun, H Tian, H Wu… - arXiv preprint arXiv …, 2020"],"snippet":"… XLM-R (Conneau et al., 2019) studies the effects of models when trained on a large scale corpus. It used 2.5T data extracted from Common Crawl (Wenzek et al., 2019) that involves 100 languages for MMLM training. Results …","url":["https://arxiv.org/pdf/2012.15674"]}
{"year":"2021","title":"Error Analysis of using BART for Multi-Document Summarization: A Study for English and German Language","authors":["T Johner, A Jana, C Biemann"],"snippet":"… For instance, single-document summarization of text in German language was done by Parida and Motlicek (2019) who utilized transformer models for abstractive summarization on two datasets — SwissText 20191 and …","url":["https://www.researchgate.net/profile/Abhik-Jana/publication/351548608_Error_Analysis_of_using_BART_for_Multi-Document_Summarization_A_Study_for_English_and_German_Language/links/609cfb0c299bf1259ed0452e/Error-Analysis-of-using-BART-for-Multi-Document-Summarization-A-Study-for-English-and-German-Language.pdf"]}
{"year":"2021","title":"Error identification for machine translation with metric embedding and attention","authors":["R Rubino, A Fujita, B Marie - Proceedings of the 2nd Workshop on Evaluation and …, 2021"],"snippet":"Abstract Quality Estimation (QE) for Machine Translation has been shown to reach relatively high accuracy in predicting sentence-level scores, relying on pretrained contextual embeddings and human-produced quality scores. However, the lack of …","url":["https://aclanthology.org/2021.eval4nlp-1.15.pdf"]}
{"year":"2021","title":"ESPnet-ST IWSLT 2021 Offline Speech Translation System","authors":["H Inaguma, B Yan, S Dalmia, P Gu, J Shi, K Duh… - arXiv preprint arXiv …, 2021"],"snippet":"… Must-C - 0.68M Must-C v2 0.74M ST-TED (cleaned) 0.40M Europarl 1.82M Commoncrawl 2.39M Paracrawl 34.37M NewsCommentary 0.37M WikiTitles 1.38M RAPID 1.63M WikiMatrix 1.57M Table 1: Corpus statistics data was …","url":["https://arxiv.org/pdf/2107.00636"]}
{"year":"2021","title":"Est-ce que vous compute? Code-switching, cultural identity, and AI","authors":["A Falbo, T LaCroix - arXiv preprint arXiv:2112.08256, 2021"],"snippet":"… of the Common Crawl dataset, resulting in approximately 14 billion tokens (though they do not provide details of how the Common Crawl … is predominantly English: around 45% of HTML pages in the Common Crawl dataset have English as their …","url":["https://arxiv.org/pdf/2112.08256"]}
{"year":"2021","title":"Establishing Trustworthiness Through Algorithmic Approaches to Qualitative Research","authors":["H Nguyen, J Ahn, A Belgrave, J Lee, L Cawelti, HE Kim… - International Conference on …, 2021","HE Kim, Y Prado, R Santagata, A Villavicencio - … 2020, Malibu, CA, USA, February 1-3 …, 2021"],"snippet":"… The model contains 300 dimensional word vectors that were trained on a vocabulary of 2 million words from web page data (Common Crawl dataset; GloVe, [21]). We then worked to cluster words together using the …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=k90XEAAAQBAJ&oi=fnd&pg=PA47&dq=commoncrawl&ots=6DvGd4cN1o&sig=FHX30JOd7Xl7m8NWbE-QfeuJbks","https://link.springer.com/chapter/10.1007/978-3-030-67788-6_4"]}
{"year":"2021","title":"Estimating the Effects of Text Genre, Image Resolution and Algorithmic Complexity needed for Sinhala Optical Character Recognition","authors":["I Anuradha, C Liyanage, R Weerasinghe - International Journal on Advances in ICT …, 2021"],"snippet":"… 2) 5million+ sentences in Sinhala common crawler: In 2019, Guzman [16] presented two monolingual corpora for Sinhala. Those were a combination of 155k+ sentences of filtered Sinhala Wikipedia and 5178k+ sentences of Sinhala common crawl …","url":["https://icter.sljol.info/articles/10.4038/icter.v14i3.7231/galley/5596/download/"]}
{"year":"2021","title":"Estimation of Imageability Ratings of English Words Using Neural Networks","authors":["VV Bochkarev, AV Savinkov, AV Shevlyakova - Mexican International Conference on …, 2021"],"snippet":"… (a set of CommonCrawl vectors). The values of the correlation coefficients obtained in [6] are shown in the last row of the table. We chose the best values from those given in [6] that were obtained also using the fastText vectors trained on the …","url":["https://link.springer.com/chapter/10.1007/978-3-030-89820-5_5"]}
{"year":"2021","title":"Evaluating and Explaining Natural Language Generation with GenX","authors":["K Duskin, S Sharma, JY Yun, E Saldanha, D Arendt - … on Data Science with Human in …, 2021"],"snippet":"… While we demonstrated utility on datasets with tens of thousands of text examples, the nearest neighbor approach used would become intractable on massive text corpora such as CommonCrawl 7. This limits GenX to …","url":["https://www.aclweb.org/anthology/2021.dash-1.12.pdf"]}
{"year":"2021","title":"Evaluating Contextualized Language Models for Hungarian","authors":["J Ács, D Lévai, D Nemeskey, A Kornai"],"snippet":"… It was trained on Webcorpus 2.0 (Nemeskey, 2020), 9- billion-token corpus compiled from the Hungarian subset of Common Crawl3. Its string identifier in Huggingface Transformers is … XLMRoBERTa was trained on 2TB of …","url":["https://hlt.bme.hu/media/pdf/huBERT_eval.pdf"]}
{"year":"2021","title":"Evaluating Document Coherence Modelling","authors":["A Shen, M Mistica, B Salehi, H Li, T Baldwin, J Qi - arXiv preprint arXiv:2103.10133, 2021"],"snippet":"Page 1. arXiv:2103.10133v1 [cs.CL] 18 Mar 2021 Evaluating Document Coherence Modelling Aili Shen ♣ , Meladel Mistica ♣ , Bahar Salehi ♣ , Hang Li ♢ , Timothy Baldwin ♣ , Jianzhong Qi ♣ ♣ The University of Melbourne …","url":["https://arxiv.org/pdf/2103.10133"]}
{"year":"2021","title":"Evaluating Off-the-Shelf Machine Listening and Natural Language Models for Automated Audio Captioning","authors":["B Weck, X Favory, K Drossos, X Serra - arXiv preprint arXiv:2110.07410, 2021"],"snippet":"… We use the publicly available model trained with subword information on the Common Crawl corpus, which contains 600B tokens and is significantly larger than the corpora used for the Glove and word2vec model [38]. We employ BERT as our …","url":["https://arxiv.org/pdf/2110.07410"]}
{"year":"2021","title":"Evaluating Sequence-to-Sequence Modelling for Dialogue State Tracking","authors":["M Tuli, S Agrawal"],"snippet":"… 3.2.1 Word embeddings. We use GloVe [21] pretrained word embeddings, as done in [9]. We use the GloVe-840B-300D embeddings: 300-dimensional embeddings trained on 840B tokens from Common Crawl. 3.2.2 Models …","url":["https://www.mathieutuli.com/docs/nmt_for_dst.pdf"]}
{"year":"2021","title":"Evaluating the Evaluation Metrics for Style Transfer: A Case Study in Multilingual Formality Transfer","authors":["E Briakou, S Agrawal, J Tetreault, M Carpuat - arXiv preprint arXiv:2110.10668, 2021"],"snippet":"While the field of style transfer (ST) has been growing rapidly, it has been hampered by a lack of standardized practices for automatic evaluation. In this paper, we evaluate leading ST automatic metrics on the oft-researched task of formality style …","url":["https://arxiv.org/pdf/2110.10668"]}
{"year":"2021","title":"Evaluating the Text-to-SQL Capabilities of Large Language Models","authors":["MVAEX TS","N Rajkumar, R Li, D Bahdanau - arXiv preprint arXiv:2204.00498, 2022"],"snippet":"We perform an empirical evaluation of Text-to-SQL capabilities of the Codex language model. We find that, without any finetuning, Codex is a strong baseline on the Spider benchmark; we also analyze the failure modes of Codex in this setting …","url":["https://arxiv.org/pdf/2204.00498","https://openreview.net/pdf?id=lYli-bAuK54"]}
{"year":"2021","title":"Evaluating the textbugger NLP Attack on","authors":["S Eicher"],"snippet":"… 1 Introduction The ubiquity of deep learning models which learn from their user-base, like social recommenders or CAPTCHA[4], or other relatively unsupervised datasets, like the Common Crawl, in practice has given rise …","url":["http://cs229.stanford.edu/proj2021spr/report2/82526938.pdf"]}
{"year":"2021","title":"Evaluating Word Embeddings with Categorical Modularity","authors":["S Casacuberta, K Halevy, DE Blasi - arXiv preprint arXiv:2106.00877, 2021"],"snippet":"… these embeddings. FastText. Monolingual embeddings for 157 languages trained on Common Crawl and Wikipedia that use CBOW with position-weights and character n-grams (Bojanowski et al., 2017). MUSE. Cross-lingual …","url":["https://arxiv.org/pdf/2106.00877"]}
{"year":"2021","title":"Evaluation and Interpretation of Word Embeddings","authors":["J Kuchár"],"snippet":"… 2.3.1 Facebook's Wikipedia and Common Crawl FastText Word Embeddings for 157 languages Produced by Grave et al. [8], the Facebook's Wikipedia and Common Crawl fastText word embeddings may be the most-used and …","url":["https://is.muni.cz/th/eycxc/Evaluation_and_interpretation_of_word_embeddings_Archive.pdf"]}
{"year":"2021","title":"Evaluation of contextual embeddings on less-resourced languages","authors":["M Ulčar, A Žagar, CS Armendariz, A Repar, S Pollak… - arXiv preprint arXiv …, 2021"],"snippet":"… The ELMoForManyLangs project (EFML) [Che et al., 2018] trained ELMo models for several languages but used relatively small datasets of 20 million words randomly sampled from the raw text released by the CONLL 2017 …","url":["https://arxiv.org/pdf/2107.10614"]}
{"year":"2021","title":"Evaluation of Neural Network Transformer Models for Named-Entity Recognition on Low-Resourced Languages","authors":["R Hanslo - 2021"],"snippet":"… LANGUAGE MODEL ARCHITECTURE XLM-Roberta (XLM-R) is a transformer-based multilingual masked language model [2]. This language model trained on 100 languages uses 2.5 TB of CommonCrawl (CC) data …","url":["https://annals-csis.org/proceedings/2021/pliks/7.pdf"]}
{"year":"2021","title":"Evaluation of Sentence Representations in Semantic Text Similarity Tasks","authors":["N Balzar Ekenbäck - 2021"],"snippet":"… 5. Transformer based encoder, using RoBERTa. For the baseline representation, word embeddings trained on common crawl data was used.[30] A few different approaches were used to construct the sentence vector representation …","url":["https://www.diva-portal.org/smash/get/diva2:1535961/FULLTEXT01.pdf"]}
{"year":"2021","title":"Evaluation of Split-and-Rephrase Output of the Knowledge Extraction Tool in the Intelligent Tutoring System","authors":["A Grubišić, B Žitko, A Gašpar, D Vasić, A Dodaj - Expert Systems with Applications, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0957417421012562"]}
{"year":"2021","title":"Everything Is All It Takes: A Multipronged Strategy for Zero-Shot Cross-Lingual Information Extraction","authors":["M Yarmohammadi, S Wu, M Marone, H Xu, S Ebner… - arXiv preprint arXiv …, 2021"],"snippet":"… We focus on the coarsest “Abstract” level, where the goal is to identify events and their agents and patients. The documents come from the news-specific portion of Common Crawl … We use the same Common Crawl corpus as XLM-R for pretraining …","url":["https://arxiv.org/pdf/2109.06798"]}
{"year":"2021","title":"Evolution of diversity and dominance of companies in online activity","authors":["PX McCarthy, X Gong, S Eghbal, DS Falster, MA Rizoiu - Plos one, 2021"],"snippet":"… We compile the linkage record over time of online websites towards domains using Common Crawl (https://commoncrawl.org/). Common Crawl is the largest open index of the web and which has been shown to represent over …","url":["https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0249993"]}
{"year":"2021","title":"Evolution of Gaussian Process kernels for machine translation post-editing effort estimation","authors":["I Roman, R Santana, A Mendiburu, JA Lozano - Annals of Mathematics and Artificial …, 2021"],"snippet":"Page 1. Annals of Mathematics and Artificial Intelligence https://doi.org/10.1007/s10472021-09751-5 Evolution of Gaussian Process kernels for machine translation post-editing effort estimation Ibai Roman1 ·Roberto Santana1 ·Alexander Mendiburu1 ·Jose A. Lozano1 …","url":["https://link.springer.com/article/10.1007/s10472-021-09751-5"]}
{"year":"2021","title":"Examination Committee.","authors":["S Colcher - 2021"],"snippet":"… evaluates the model on a German Wiki data 9 using data extracted from Common Crawl 10 as synthetic summaries … it delivers a good value on aiding the overall performance assessment. 9https://www.swisstext.org/ 10http://commoncrawl.org/ Page 18 …","url":["https://www.maxwell.vrac.puc-rio.br/53550/53550.PDF"]}
{"year":"2021","title":"Examining Media Bias in the 2020 Democratic Primary","authors":["D Färber - 2021"],"snippet":"Page 1. Examining Media Bias in the 2020 Democratic Primary by Extraction of Entity-Relationships from News Articles Master Thesis submitted by Daniel Färber at University of Konstanz Faculty of Politics, Law and Economics …","url":["https://www.researchgate.net/profile/Daniel-Faerber/publication/354219152_Examining_Media_Bias_in_the_2020_Democratic_Primary_by_Extraction_of_Entity-Relationships_from_News_Articles/links/612cd83f2b40ec7d8bd2405b/Examining-Media-Bias-in-the-2020-Democratic-Primary-by-Extraction-of-Entity-Relationships-from-News-Articles.pdf"]}
{"year":"2021","title":"Examining the Effects of Preprocessing on the Detection of Offensive Language in German Tweets","authors":["S Reimann, D Dakota"],"snippet":"… dings. We ultimately settled on on the default 300 dimensional fastText German embeddings (Grave et al., 2018), trained on the German CommonCrawl and Wikpedia, as they yielded the most stable performance. We also …","url":["https://konvens2021.phil.hhu.de/wp-content/uploads/2021/09/2021.KONVENS-1.14.pdf"]}
{"year":"2021","title":"Executive function & semantic memory impairments in Alzheimer's disease—investigating the decline of executive function and semantic memory in Alzheimer's …","authors":["JM Tröger - 2021"],"snippet":"Alzheimer’s Disease (AD) has a huge impact on an ever-aging society in highly developed industrialized countries such as the EU member states: according to the World Alzheimer’s Association the number one risk factor for AD is age. AD patients …","url":["https://publikationen.sulb.uni-saarland.de/bitstream/20.500.11880/31994/1/Troeger-Executive_Function_vs._Semantic_Memory_in_AD-v3.4.pdf"]}
{"year":"2021","title":"Experience and prediction: a metric of hardness for a novel litmus test","authors":["N Isaak, L Michael - Journal of Logic and Computation, 2021"],"snippet":"Abstract. In the past decade, the Winograd schema challenge (WSC) has become a central aspect of the research community as a novel litmus test. Consequently, th.","url":["https://academic.oup.com/logcom/advance-article-abstract/doi/10.1093/logcom/exab005/6129485"]}
{"year":"2021","title":"Experience Report: Deep Learning-based System Log Analysis for Anomaly Detection","authors":["Z Chen, J Liu, W Gu, Y Su, MR Lyu - arXiv preprint arXiv:2107.05908, 2021"],"snippet":"Page 1. Experience Report: Deep Learning-based System Log Analysis for Anomaly Detection Zhuangbin Chen, Jinyang Liu, Wenwei Gu The Chinese University of Hong Kong Hong Kong, China {zbchen,jyliu,wwgu}@cse.cuhk.edu.hk …","url":["https://arxiv.org/pdf/2107.05908"]}
{"year":"2021","title":"Experimental Evaluation of Deep Learning models for Marathi Text Classification","authors":["A Kulkarni, M Mandhane, M Likhitkar, G Kshirsagar… - arXiv preprint arXiv …, 2021"],"snippet":"… It was created by processing January-December 2018 Commoncrawl snapshots … OSCAR: Open Super-large Crawled AL- MAnaCH coRpus (OSCAR) is obtained by filtering and language classifying the Common Crawl corpus (Suárezetal., 2019) …","url":["https://arxiv.org/pdf/2101.04899"]}
{"year":"2021","title":"Experiments with adversarial attacks on text genres","authors":["M Lepekhin, S Sharoff - arXiv preprint arXiv:2107.02246, 2021"],"snippet":"… It has the same architecture, but uses bigger and more genre diverse corpora and an updated pre-training procedure. In addition, XLM-RoBERTa is a multilingual model trained on Common Crawl data in comparison to multilingual BERT only trained on Wikipedia …","url":["https://arxiv.org/pdf/2107.02246"]}
{"year":"2021","title":"ExpertRank: A Multi-level Coarse-grained Expert-based Listwise Ranking Loss","authors":["Z Chen, C Eickhoff - arXiv preprint arXiv:2107.13752, 2021"],"snippet":"… For KNRM, ConvKNRM and MatchPyramid, GloVe word embeddings with 300 dimensions [25], trained on the Common Crawl dataset2, are used for vocabulary initialization … 1https://www.tensor ow.org/ 2https://commoncrawl …","url":["https://arxiv.org/pdf/2107.13752"]}
{"year":"2021","title":"Explainable Prediction of Text Complexity: The Missing Preliminaries for Text Simplification","authors":["C Gârbacea, M Guo, S Carton, Q Mei"],"snippet":"… all positions without data corruption. We use the 12 layer XLNeT base pre-trained model on the English Wikipedia, the Books corpus (similar to BERT), Giga5, ClueWeb 2012-B, and Common Crawl. 4.1.2 Evaluation Metric We …","url":["https://aclanthology.org/2021.acl-long.88.pdf"]}
{"year":"2021","title":"Exploiting Linguistic Information from Nepali Transcripts for Early Detection of Alzheimer's Disease using Natural Language Processing and Machine Learning …","authors":["S Adhikari, S Thapa, U Naseem, P Singh, H Huo… - International Journal of …, 2021"],"snippet":"Alzheimer's disease (AD) is considered as progressing brain disease, which can be slowed down with the early detection and proper treatment by identifying the early symptoms. Language change serves as an early sign that a patient's cognitive …","url":["https://www.sciencedirect.com/science/article/pii/S1071581921001798"]}
{"year":"2021","title":"Exploiting semantic similarity models to automate transfer credit assessment in academic mobility","authors":["D Chandrasekaran - 2021"],"snippet":"Page 1. Exploiting semantic similarity models to automate transfer credit assessment in academic mobility by Dhivya Chandrasekaran A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer Science in the …","url":["https://lurepository.lakeheadu.ca/bitstream/handle/2453/4779/ChandrasekaranD2021m-1a.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Exploring Conditional Text Generation for Aspect-Based Sentiment Analysis","authors":["SUS Chebolu, F Dernoncourt, N Lipka, T Solorio - arXiv preprint arXiv:2110.02334, 2021"],"snippet":"… (2019). The encoder utilizes a denoising goal similar to BERT, while the decoder tries to recreate the original sequence (autoencoder) token by token using the preceding (uncorrupted) tokens and the encoder output. T5 uses common crawl web extracted text …","url":["https://arxiv.org/pdf/2110.02334"]}
{"year":"2021","title":"Exploring deep learning methods for recognizing rare diseases and their clinical manifestations from texts","authors":["I Segura-Bedmar, D Camino-Perdonas… - arXiv preprint arXiv …, 2021"],"snippet":"… The model contains word embeddings of di- mension 300 for 3 million words. • GloVe [54], a pre-trained word embedding model trained using Common Crawl, an open repository of web crawl data. The model contains …","url":["https://arxiv.org/pdf/2109.00343"]}
{"year":"2021","title":"Exploring intrinsic information content models for addressing the issues of traditional semantic measures to evaluate verb similarity","authors":["MKS Prasad, P Sharma - Computer Speech & Language, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0885230821000838"]}
{"year":"2021","title":"Exploring Monolingual Data for Neural Machine Translation with Knowledge Distillation","authors":["AF Aji, K Heafield - arXiv preprint arXiv:2012.15455, 2020"],"snippet":"… Out- of-domain back-translation data (CommonCrawl– mixed, Open Subtitle–conversation) damages the performance, however … Though, CommonCrawl or Opensub1 conversational monolingual data is still better than not us- ing any monolingual data at all …","url":["https://arxiv.org/pdf/2012.15455"]}
{"year":"2021","title":"Exploring Pre-Trained Transformers and Bilingual Transfer Learning for Arabic Coreference Resolution","authors":["B Min - Proceedings of the Fourth Workshop on Computational …, 2021"],"snippet":"Abstract In this paper, we develop bilingual transfer learning approaches to improve Arabic coreference resolution by leveraging additional English annotation via bilingual or multilingual pre-trained transformers. We show that bilingual transfer …","url":["https://aclanthology.org/2021.crac-1.10.pdf"]}
{"year":"2021","title":"Exploring the Possibilities of Applying Transfer Learning Methods for Natural Language Processing in Software Development","authors":["W Ding"],"snippet":"Page 1. DEPARTMENT OF INFORMATICS TECHNISCHE UNIVERSITÄT MÜNCHEN Master's Thesis in Informatics: Data Engineering and Analytics Exploring the Possibilities of Applying Transfer Learning Methods for Natural …","url":["https://wwwmatthes.in.tum.de/file/1d56vodcuu5ou/Sebis-Public-Website/-/Master-Thesis-Wei-Ding/Masterthesis_WeiDing.pdf"]}
{"year":"2021","title":"Exploring the Predictive Power of News and Neural Machine Learning Models for Economic Forecasting","authors":["V Bitetta, I Bordino, A Ferretti, F Gullo, G Ponti… - Mining Data for Financial …"],"snippet":"… rule-based procedure that builds on the linguistic features of the spaCy Python library 5 . The NLP pipeline relies on the en_core_web_lg model of spaCy 6 , an English multi-task Convolutional Neural Network trained on OntoNotes …","url":["https://europepmc.org/article/pmc/pmc7808169"]}
{"year":"2021","title":"Exploring Transformers in Emotion Recognition: a comparison of BERT, DistillBERT, RoBERTa, XLNet and ELECTRA","authors":["D Cortiz - arXiv preprint arXiv:2104.02041, 2021"],"snippet":"… tokens but in random order. XLNet was trained with over 130 GB of textual data. In addition to BERT's two datasets, they included three more corpora: Giga5, ClueWeb, and Common Crawl. XLNet outperformed BERT on 20 …","url":["https://arxiv.org/pdf/2104.02041"]}
{"year":"2021","title":"Exploring Unsupervised Pretraining Objectives for Machine Translation","authors":["C Baziotis, I Titov, A Birch, B Haddow - arXiv preprint arXiv:2106.05634, 2021"],"snippet":"… (2019). The (pretraining) monolingual data contain 5M sentences from Common Crawl and Wikipedia per language, while the parallel data are approximately 600K sentences from the Bible, Open Subtitles, GNOME/KDE/Ubuntu, and Paracrawl …","url":["https://arxiv.org/pdf/2106.05634"]}
{"year":"2021","title":"Exploring Zero-shot Cross-lingual Aspect-based Sentiment Analysis using Pre-trained Multilingual Language Models","authors":["KTK Phan, DN Hao, D Van Thin, NLT Nguyen - 2021 International Conference on …, 2021"],"snippet":"… The mBERT model is trained on the Wikipedia data of 104 languages, while the XLM-R model is trained on the CommonCrawl data in 100 … In detail, the XLM-R model is trained on the combination of Common Crawl with Wikipedia data, while …","url":["https://ieeexplore.ieee.org/abstract/document/9585242/"]}
{"year":"2021","title":"Exploring Zero-Shot Emotion Recognition in Speech Using Semantic-Embedding Prototypes","authors":["X Xu, J Deng, N Cummins, Z Zhang, L Zhao… - IEEE Transactions on …, 2021"],"snippet":"Page 1. 1520-9210 (c) 2021 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This …","url":["https://ieeexplore.ieee.org/abstract/document/9449953/"]}
{"year":"2021","title":"Exposure of occupations to technologies of the fourth industrial revolution","authors":["B Meindl, MR Frank, J Mendonça - arXiv preprint arXiv:2110.13317, 2021"],"snippet":"… The common crawl corpus comprises texts from various sources available on the web, and therefore the embeddings represent various aspects of language. However, the structure and words used in patent data are more technical than standard …","url":["https://arxiv.org/pdf/2110.13317"]}
{"year":"2021","title":"EXPRESS: Attribute Sentiment Scoring with Online Text Reviews: Accounting for Language Structure and Missing Attributes","authors":["I Chakraborty, M Kim, K Sudhir - Journal of Marketing Research, 2021"],"snippet":"Page 1. Peer Review Version Attribute Sentiment Scoring with Online Text Reviews: Accounting for Language Structure and Missing Attributes Journal: Journal of Marketing Research Manuscript ID JMR.19.0192.R3 Manuscript …","url":["https://journals.sagepub.com/doi/abs/10.1177/00222437211052500"]}
{"year":"2021","title":"Extending Neural Temporal Tagging Systems with External Knowledge","authors":["M Pömsl"],"snippet":"Page 1. Extending Neural Temporal Tagging Systems with External Knowledge Martin Pömsl Dr. phil. Tobias Thelen Institute of Cognitive Science, Osnabrück M. Sc. Luzian Hahn Fraunhofer Institute for Integrated Circuits, Erlangen Osnabrück University …","url":["https://mpoemsl.github.io/documents/bachelor_thesis_mpoemsl.pdf"]}
{"year":"2021","title":"Extracting ESG data from business documents","authors":["J Van Der Elst, S Nijssen"],"snippet":"… Common Crawl (Want to use our data? – Common Crawl 2021) for example is an open repository of web crawl data consisting of web page data, metadata extracts and text extracts. It is obtained by scraping web pages and ignoring HTML markup …","url":["https://dial.uclouvain.be/downloader/downloader.php?pid=thesis%3A30732&datastream=PDF_01&cover=cover-mem"]}
{"year":"2021","title":"Extraction and Analysis of Stories: Natural Language Processing for Software Engineering.","authors":["H Guo - 2021"],"snippet":"Page 1. ABSTRACT GUO, HUI. Extraction and Analysis of Stories: Natural Language Processing for Software Engineering. (Under the direction of Dr. Munindar P. Singh.) Modern software systems are deployed in sociotechnical …","url":["https://repository.lib.ncsu.edu/bitstream/handle/1840.20/38875/etd.pdf?sequence=1"]}
{"year":"2021","title":"Extremely low-resource neural machine translation for Asian languages","authors":["R Rubino, B Marie, R Dabre, A Fujita, M Utiyama… - Machine Translation, 2021"],"snippet":"This paper presents a set of effective approaches to handle extremely low-resource language pairs for self-attention based neural machine translation (NMT).","url":["https://link.springer.com/article/10.1007/s10590-020-09258-6"]}
{"year":"2021","title":"Facebook AI WMT21 News Translation Task Submission","authors":["C Tran, S Bhosale, J Cross, P Koehn, S Edunov, A Fan - arXiv preprint arXiv …, 2021"],"snippet":"… For each language, we train an n-gram language model (Heafield, 2011) on all available news-domain data (Newscrawl) and a n-gram language model on a similarly sized sample from general-domain data (Commoncrawl) …","url":["https://arxiv.org/pdf/2108.03265"]}
{"year":"2021","title":"Fact-Checking of Claims from the English Wikipedia Using Evidence in the Wild","authors":["A Sathe - 2021"],"snippet":"… architectures. These advancements enable taking advantage of vast amounts of unsupervised text data obtained using the internet, via sources such as Wikipedia and Common Crawl. Using these large corpora, researchers have demonstrated …","url":["https://scholarship.richmond.edu/cgi/viewcontent.cgi?article=2562&context=honors-theses"]}
{"year":"2021","title":"Factorising Meaning and Form for Intent-Preserving Paraphrasing","authors":["T Hosking, M Lapata - arXiv preprint arXiv:2105.15053, 2021"],"snippet":"… English. Although this system was trained on high volumes of data (including Common Crawl), the training data contains relatively few questions, and we would not expect it to perform well in the domain under consideration …","url":["https://arxiv.org/pdf/2105.15053"]}
{"year":"2021","title":"Fast and Compact Set Intersection through Recursive Universe Partitioning","authors":["GE Pibiri"],"snippet":"… CCNews is a dataset of news freely available from CommonCrawl; precisely, it consists of the news appeared from 09/01/16 to 30/03/18. Identifiers were assigned to documents according to the lexicographic order of their URLs …","url":["http://pages.di.unipi.it/pibiri/papers/DCC21.pdf"]}
{"year":"2021","title":"Fast Extraction of Word Embedding from Q-contexts","authors":["J Kong, W Li, Z Liu, B Liao, J Qiu, CY Hsieh, Y Cai… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Fast Extraction of Word Embedding from Q-contexts Junsheng Kong ∗† School of Software Engineering, South China University of Technology China sescut_kongjunsheng@mail.scut.edu.cn Weizhao Li ∗† School …","url":["https://arxiv.org/pdf/2109.07084"]}
{"year":"2021","title":"FAST: A carefully sampled and cognitively motivated dataset for distributional semantic evaluation","authors":["S Evert, G Lapesa"],"snippet":"… Embeddings: word2vec (Mikolov et al., 2013) trained on 100G tokens of Google News data,10 GloVe (Pennington et al., 2014) trained on 6G to- kens of wikipedia + newspapers or 42G tokens of Common Crawl (CC) Web …","url":["https://files.de-1.osf.io/v1/resources/cd8ar/providers/osfstorage/613cdcae079a0a027601be5b?action=download&direct&version=1"]}
{"year":"2021","title":"Faster Re-translation Using Non-Autoregressive Model For Simultaneous Neural Machine Translation","authors":["H Han, S Indurthi, MA Zaidi, NK Lakumarapu, B Lee… - arXiv preprint arXiv …, 2020"],"snippet":"… We choose the best system based on the MuST-C dev set and report the results on the MuST-C tst-COMMON test set. The WMT19 dataset further consists of Europarl v9, ParaCrawl v3, Common Crawl, News Commentary …","url":["https://arxiv.org/pdf/2012.14681"]}
{"year":"2021","title":"fastText-based methods for Emotion Identification in Russian Internet Discourse","authors":["A Babii, M Kazyulina, A Malafeev - 13th ACM Web Science Conference 2021, 2021"],"snippet":"… 0.8. The only exception is the model based on the GRU architecture and pre-trained on the Common Crawl. That … 0.81. The lowest performance was shown by the CNN model pre-trained on the Common Crawl embeddings. We …","url":["https://dl.acm.org/doi/abs/10.1145/3447535.3462499"]}
{"year":"2021","title":"FEEL-IT: Emotion and Sentiment Classification for the Italian Language","authors":["F Bianchi, D Nozza, D Hovy - Proceedings of the Eleventh Workshop on …, 2021"],"snippet":"… In this paper, we use the Italian BERT model UmBERTo trained on Commoncrawl ITA.4 As the first experimental condition, we fine-tune the UmBERTo model for the task of emotion classification with the considered training data (UmBERTo-FT) …","url":["https://www.aclweb.org/anthology/2021.wassa-1.8.pdf"]}
{"year":"2021","title":"Felix Hamborg, Philipp Meschenmoser, Moritz Schubotz","authors":["P Scharpf, B Gipp - Wahrheit und Fake im postfaktisch-digitalen Zeitalter"],"snippet":"… These filter criteria are passed to news-please, which subsequently searches for relevant news articles in the Common Crawl News … 2 News-please currently accesses the raw archive provided by the Common Crawl project. To speed up the …","url":["https://link.springer.com/content/pdf/10.1007/978-3-658-32957-0.pdf#page=164"]}
{"year":"2021","title":"Feminist Curation of Text for Data-centric AI","authors":["M Bartl, S Leavy"],"snippet":"Abstract Language models are becoming increasingly central to artificial intelligence through their use in online search, recommendation engines and language generation technologies. However, concepts of gender can be deeply embedded in …","url":["https://datacentricai.org/papers/79_CameraReady_DCAI_Workshop_NeurIPS_2021_final.pdf"]}
{"year":"2021","title":"Few-shot Learning with Multilingual Language Models","authors":["XV Lin, T Mihaylov, M Artetxe, T Wang, S Chen… - arXiv preprint arXiv …, 2021"],"snippet":"… Our models are trained on a static multilingual corpus extracted from CommonCrawl, with English text comprising 32.6% of the total number … As such we also note the potential difference in genres between CommonCrawl and the genres used in GPT-3 …","url":["https://arxiv.org/pdf/2112.10668"]}
{"year":"2021","title":"FH-SWF SG at GermEval 2021: Using Transformer-Based Language Models to Identify Toxic, Engaging, & Fact-Claiming Comments","authors":["C Gawron, S Schmidt - arXiv preprint arXiv:2109.02966, 2021"],"snippet":"… Philipp Reissel and Philip May have published both a German ELECTRA model (Reissel and May, 2020) and a “German colossal, cleaned Common Crawl corpus” (GC4) (Reissel and May, 2021) with about 540 GB of German …","url":["https://arxiv.org/pdf/2109.02966"]}
{"year":"2021","title":"Findings of the WMT 2021 Biomedical Translation Shared Task: Summaries of Animal Experiments as New Test Set","authors":["L Yeganova, D Wiemann, M Neves, F Vezzani, A Siu… - Sixth Conference on …, 2021"],"snippet":"In the sixth edition of the WMT Biomedical Task, we addressed a total of eight language pairs, namely English/German, English/French, English/Spanish, English/Portuguese, English/Chinese, English/Russian, English/Italian, and English/Basque. Further, our …","url":["https://hal.archives-ouvertes.fr/hal-03435096/document"]}
{"year":"2021","title":"Fine and Coarse Granular Argument Classification before Clustering","authors":["L Dumani, T Wiesenfeldt, R Schenkel - Proceedings of the 30th ACM International …, 2021"],"snippet":"… This dataset consists of eight topics each with 1,000 usergenerated sentences originating from the Common Crawl archive.A key difference to our dataset, besides presumably different appearing frames as well as topics being used as queries, is …","url":["https://dl.acm.org/doi/abs/10.1145/3459637.3482431"]}
{"year":"2021","title":"Fine-tuning Distributional Semantic Models for Closely-Related Languages","authors":["K Bhatia, D Aggarwal, A Vaidya - Proceedings of the Eighth Workshop on NLP for …, 2021"],"snippet":"… We further compare our fine-tuned models with two pre-trained word embedding models, namely fastText (FT-WC), trained on Common Crawl and Wikipedia (Grave et al., 2018) and IndicFastText (I- FT) (Kakwani et al., 2020) …","url":["https://www.aclweb.org/anthology/2021.vardial-1.7.pdf"]}
{"year":"2021","title":"FinMatcher at FinSim-2: Hypernym Detection in the Financial Services Domain using Knowledge Graphs","authors":["J Portisch, M Hladik, H Paulheim - arXiv preprint arXiv:2103.01576, 2021"],"snippet":"… DBpedia [10] – but instead on the whole Web: The dataset consists of hypernymy relations extracted from the Common Crawl5, a … 3see https://www.wikidata.org/wiki/Q1881843 4see https://query.wikidata.org …","url":["https://arxiv.org/pdf/2103.01576"]}
{"year":"2021","title":"Focused Attention Improves Document-Grounded Generation","authors":["S Prabhumoye, K Hashimoto, Y Zhou, AW Black… - arXiv preprint arXiv …, 2021"],"snippet":"… 4.1 Wikipedia Update Generation This task involves generating an update for Wikipedia context given a news article (Prabhumoye et al., 2019). The dataset was collected by parsing Wikipedia articles and Common Crawl for news articles …","url":["https://arxiv.org/pdf/2104.12714"]}
{"year":"2021","title":"FoodChem: A food-chemical relation extraction model","authors":["G Cenikj, BK Seljak, T Eftimov - arXiv preprint arXiv:2110.02019, 2021"],"snippet":"… Apart from the datasets used for pre-training the original BERT model, 3 additional sources are involved in the pre-training of RoBERTa: the OpenWebText corpus [32], the Stories subset from the Common Crawl dataset [33] and the CommonCrawl News dataset [34] …","url":["https://arxiv.org/pdf/2110.02019"]}
{"year":"2021","title":"Four dimensions characterize attributions from faces using a representative set of English trait words","authors":["C Lin, U Keles, R Adolphs - Nature Communications, 2021"],"snippet":"People readily (but often inaccurately) attribute traits to others based on faces. While the details of attributions depend on the language available to describe social traits, psychological theories argue that two or three dimensions (such …","url":["https://www.nature.com/articles/s41467-021-25500-y"]}
{"year":"2021","title":"From Argument Search to Argumentative Dialogue: A Topic-independent Approach to Argument Acquisition for Dialogue Systems","authors":["N Rach, C Schindler, I Feustel, J Daxenberger…"],"snippet":"… query. The engine utilizes a web crawl from the year 2016 based on CommonCrawl 2 to retrieve relevant documents and subsequently classify sentences in the documents as either pro, con or no argument (Stab et al., 2018) …","url":["https://sigdial.org/sites/default/files/workshops/conference22/Proceedings/pdf/2021.sigdial-1.39.pdf"]}
{"year":"2021","title":"From OpenAPI Fragments to API Pattern Primitives and Design Smells","authors":["S Serbout, C Pautasso, U Zdun, O Zimmermann - 2021"],"snippet":"In the past few years, the OpenAPI Specification (OAS) has emerged as a standard description language for accurately modeling Web APIs. Today, thousands of OpenAPI descriptions can be found by mining open source repositories. In this paper …","url":["http://eprints.cs.univie.ac.at/7145/1/main.pdf"]}
{"year":"2021","title":"gaBERT--an Irish Language Model","authors":["J Barry, J Wagner, L Cassidy, A Cowap, T Lynn… - arXiv preprint arXiv …, 2021"],"snippet":"… texts, and official documents. OSCAR The unshuffled Irish portion of the OS- CAR Corpus (Ortiz Suárez et al., 2019) which is a collection of documents from CommonCrawl and sorted by language ID. ParaCrawl The Irish portion …","url":["https://arxiv.org/pdf/2107.12930"]}
{"year":"2021","title":"Gaze-based Multimodal Meaning Recovery for Noisy/Complex Environments","authors":["Ö Alaçam, G Malhotra, E Ruppert, C Biemann - Proceedings of the 2021 International …, 2021"],"snippet":"… In addition, we also utilized xlm-roberta-base masked language model, which is a large multi-lingual language model, trained on 2.5 TB of filtered CommonCrawl data [8]. Unlike multilingual BERT and RoBERTa, XLM-RoBERTa can infer the language of …","url":["https://dl.acm.org/doi/abs/10.1145/3462244.3481002"]}
{"year":"2021","title":"Gender Bias in English and German Children's Literature: A Computational Analysis Using Word Embeddings","authors":["D Geißler - 2021"],"snippet":"Page 1. Gender Bias in English and German Children's Literature: A Computational Analysis Using Word Embeddings MASTER THESIS Dominique Geißler Master Interaction Technology Faculty of Electrical Engineering, Mathematics …","url":["http://essay.utwente.nl/88570/1/Geissler_MA_EEMCS.pdf"]}
{"year":"2021","title":"Gender Stereotypes in Natural Language: Word Embeddings Show Robust Consistency Across Child and Adult Language Corpora of More Than 65 Million Words","authors":["TES Charlesworth, V Yang, TC Mann, B Kurdi… - Psychological Science, 2020"],"snippet":"… In this case, lemmatization is particularly helpful because the target corpora are smaller than other natural-language corpora, such as the Common Crawl corpus (which we use for validation as described below) with more than 600 billion word tokens (Mikolov, Grave …","url":["https://journals.sagepub.com/doi/abs/10.1177/0956797620963619"]}
{"year":"2021","title":"Generalization of cyberbullying traces","authors":["MA Larochelle - 2020"],"snippet":"Page 1. © Marc-André Larochelle, 2020 Generalization of Cyberbullying Traces Mémoire Marc-André Larochelle Maîtrise en informatique - avec mémoire Maître ès sciences (M. Sc.) Québec, Canada Page 2. Generalization of Cyberbullying Traces Mémoire …","url":["https://corpus.ulaval.ca/jspui/bitstream/20.500.11794/67899/1/36823.pdf"]}
{"year":"2021","title":"Generalized zero-shot object recognition withoutclass-attribute relations","authors":["M Er - 2021"],"snippet":"Page 1. GENERALIZED ZERO-SHOT OBJECT RECOGNITION WITHOUT CLASS-ATTRIBUTE RELATIONS A THESIS SUBMITTED TO THE GRADUATE SCHOOL OF NATURAL AND APPLIED SCIENCES OF MIDDLE EAST …","url":["https://open.metu.edu.tr/bitstream/handle/11511/89793/12626239.pdf"]}
{"year":"2021","title":"Generating Sense Inventories for Ambiguous Arabic Words","authors":["M Alian, A Awajan"],"snippet":"… 4.3. Fasttext Pre-Trained Embeddings Arabic Fasttext embeddings are provided by Grave et al. [12]. These embeddings result from training on Wikipedia and Common Crawl corpus. They use an extension of the Fasttext model with subword information …","url":["https://www.researchgate.net/profile/Marwah-Alian/publication/351762403_Generating_Sense_Inventories_for_Ambiguous_Arabic_Words/links/60b9d6f192851cb13d7442ab/Generating-Sense-Inventories-for-Ambiguous-Arabic-Words.pdf"]}
{"year":"2021","title":"Generative Adversarial Networks based on Mixed-Attentions for Citation Intent Classification in Scientific Publications","authors":["YS Wang, CY Chen, LH Lee - Proceedings of the 33rd Conference on Computational …, 2021"],"snippet":"Abstract We propose the mixed-attention-based Generative Adversarial Network (named maGAN), and apply it for citation intent classification in scientific publication. We select domain-specific training data, propose a mixed-attention mechanism, and …","url":["https://aclanthology.org/2021.rocling-1.36.pdf"]}
{"year":"2021","title":"Generative Chatbot Framework for Cybergrooming Prevention","authors":["P Wang - 2021"],"snippet":"Cybergrooming refers to the crime of establishing personal close relationships with potential victims, commonly teens, for the purpose of sexual exploitation or abuse via online social media platforms. Cybergrooming has been recognized as a serious …","url":["https://vtechworks.lib.vt.edu/bitstream/handle/10919/107136/Wang_P_T_2021.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Generative Question Answering in a Low-Resource Setting","authors":["L Isotalo"],"snippet":"… While the monolingual T5 was pre-trained on the C4 data set which consists of 750GB of English text from Common Crawl, mT5 has been pre-trained on the data likewise collected from Common Crawl but in over 100 languages …","url":["https://dke.maastrichtuniversity.nl/jan.niehues/wp-content/uploads/2021/07/Isotalo-Thesis.pdf"]}
{"year":"2021","title":"Generic Key Value Extractions from Emails","authors":["R Gupta - International Conference on Big Data Analytics, 2020"],"snippet":"… sentences. Authors of [14] describe a machine learning-based model trained on common-crawl data with multi-task optimizations to represent each keyword phrase using a 500-dimension vector of real numbers. Cosine similarity …","url":["https://link.springer.com/chapter/10.1007/978-3-030-66665-1_13"]}
{"year":"2021","title":"GenSumm: A Joint Framework for Multi-task Tweet Classification and Summarization using Sentiment Analysis and Generative Modelling","authors":["D Bansal, R Grover, N Saini, S Saha - IEEE Transactions on Affective Computing, 2021"],"snippet":"Social media platforms like Twitter act as a medium for communication among people, government agencies, NGOs, and other relief providing agencies in widespread humanitarian havoc during a disaster outbreak when other …","url":["https://ieeexplore.ieee.org/abstract/document/9629222/"]}
{"year":"2021","title":"Geoscience Language Processing for Exploration","authors":["H Denli, HA Chughtai, B Hughes, R Gistri, P Xu - Abu Dhabi International Petroleum …, 2021"],"snippet":"Deep learning has recently been providing step-change capabilities, particularly using transformer models, for natural language processing applications such as question answering, query-based summarization, and language translation for …","url":["https://onepetro.org/SPEADIP/proceedings-abstract/21ADIP/3-21ADIP/D031S102R003/474196"]}
{"year":"2021","title":"GeoVectors: A Linked Open Corpus of OpenStreetMap Embeddings on World Scale","authors":["N Tempelmeier, S Gottschalk, E Demidova - arXiv preprint arXiv:2108.13092, 2021"],"snippet":"… them. Training: Pre-trained word vectors are available at the fastText website6. As most of the OSM keys are in English, we chose the 300dimensional English word vectors trained on the Common Crawl, and Wikipedia [10] …","url":["https://arxiv.org/pdf/2108.13092"]}
{"year":"2021","title":"German Speech Recognition System using DeepSpeech","authors":["J Xu, K Matta, S Islam, A Nürnberger - … of the 4th International Conference on Natural …, 2020"],"snippet":"… In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 4774–4778. [5] Common crawl 2020. Language model. http://commoncrawl org/. [6] DeepSpeech Deutsch 2020. German dataset of Mozilla Common voice …","url":["https://dl.acm.org/doi/abs/10.1145/3443279.3443313"]}
{"year":"2021","title":"GFST: Gender-Filtered Self-Training for More Accurate Gender in Translation","authors":["PK Choubey, A Currey, P Mathur, G Dinu…"],"snippet":"Page 1. GFST: Gender-Filtered Self-Training for More Accurate Gender in Translation Prafulla Kumar Choubey∗† Texas A&M University prafulla.choubey@tamu.edu Anna Currey∗, Prashant Mathur, Georgiana Dinu Amazon …","url":["https://assets.amazon.science/db/c9/8f55d40b417ebc2266c4fb81d90f/gfst-gender-filtered-self-training-for-more-accurate-gender-in-translation.pdf"]}
{"year":"2021","title":"GitTables: A Large-Scale Corpus of Relational Tables","authors":["M Hulsebos, Ç Demiralp, P Groth - 2021","MHS Computing, ÇDS Computing"],"snippet":"… We use the character-level n-gram FastText model pretrained on the Common Crawl corpus. We experimented with contextual models like BERT [13] and the Universal Sentence Encoder [8] , but found that these models yield worse or at best …","url":["https://ar5iv.labs.arxiv.org/pdf/2106.07258.pdf","https://openreview.net/pdf?id=yYQuqGcxFvb"]}
{"year":"2021","title":"Going Full-TILT Boogie on Document Understanding with Text-Image-Layout Transformer","authors":["R Powalski, Ł Borchmann, D Jurkiewicz, T Dwojak… - arXiv preprint arXiv …, 2021"],"snippet":"… Library ∗ , and PDF files from Common Crawl. Then, a T5-like masked language model pretraining objective is used, but in a salient span masking scheme, ie, named entities are preferred rather than random tokens [47,15] …","url":["https://arxiv.org/pdf/2102.09550"]}
{"year":"2021","title":"Golos: Russian Dataset for Speech Research","authors":["N Karpov, A Denisenko, F Minkin - arXiv preprint arXiv:2106.10161, 2021"],"snippet":"… language. Additionally, we built a 3-gram language model on the open Common Crawl dataset and merged it with the train set transcriptions. Using … 8. 7https:// commoncrawl.org 8https://github.com/sberdevices/golos Page 5. 8 …","url":["https://arxiv.org/pdf/2106.10161"]}
{"year":"2021","title":"Governments and the Net: Defense, Control, and Trust in the Fifth Domain","authors":["L Kawerau - 2021"],"snippet":"… The by far largest data source is the Common Crawl archive. Every four weeks, Common Crawl generates a new dataset that contains content and metadata of billions of webpages based on following the links on a large sample of websites …","url":["http://kops.uni-konstanz.de/bitstream/handle/123456789/55972/Kawerau_2-176y1g4fs6pvh0.pdf?sequence=5&isAllowed=y"]}
{"year":"2021","title":"GPT Perdetry Test: Generating new meanings for new words","authors":["N Malkin, S Lanka, P Goel, S Rao, N Jojic - Proceedings of the 2021 Conference of …, 2021"],"snippet":"… Most of the words were known by none of them. 36 of the 74 words do not appear in the top 2 million words of the Common Crawl corpus, according to the GloVe embedding matrix (Pennington et al., 2014), and the median rank of the other words is 656565 …","url":["https://www.aclweb.org/anthology/2021.naacl-main.439.pdf"]}
{"year":"2021","title":"GPT3-to-plan: Extracting plans from text using GPT-3","authors":["A Olmo, S Sreedharan, S Kambhampati - arXiv preprint arXiv:2106.07131, 2021"],"snippet":"… 2020) is the latest version of the GPT models developed by OpenAI1. A 175 billion parameter autoregressive language model with 96 layers trained on a 560GB+ web corpora (Common Crawl2 and WebText2 (Gokaslan …","url":["https://arxiv.org/pdf/2106.07131"]}
{"year":"2021","title":"Grammar Index by Induced Suffix Sorting","authors":["H Bannai, M Takeda - String Processing and Information Retrieval: 28th …","T Akagi, D Köppl, Y Nakashima, S Inenaga, H Bannai… - arXiv preprint arXiv …, 2021"],"snippet":"… Sizes are in megabytes [MB]. Dataset commoncrawl dna einstein. de Input size GCIS-nep GCIS-uni ESP-index FM-index r-index 221. 180 220. 119 138 … com/nicolaprezza/r-index, respectively. To save space, we renamed …","url":["https://arxiv.org/pdf/2105.13744","https://books.google.de/books?hl=en&lr=lang_en&id=gEdFEAAAQBAJ&oi=fnd&pg=PA85&dq=commoncrawl&ots=0CKdrg1l4k&sig=X0BiUyMEx7vsbj3O2GcRTMn6Cms"]}
{"year":"2021","title":"Graph-based KB and Text Fusion Interaction Network for Open Domain Question Answering","authors":["Y Ding, Y Rao, F Yang - 2021 International Joint Conference on Neural …, 2021"],"snippet":"… C. Implementation Details Through our experiments, the 300-dimension GloVe embedding trained on the Common Crawl corpus is used to initialize the word embedding. The hidden dimension of LSTM and the dimension of entity embedding are both 100 …","url":["https://ieeexplore.ieee.org/abstract/document/9534439/"]}
{"year":"2021","title":"Grounding Linguistic Commands to Navigable Regions","authors":["N Rufus, K Jain, UKR Nair, V Gandhi, KM Krishna - 2021"],"snippet":"… Input images are resized to 448 × 448 spatial resolution. We use 300d GloVe embeddings pre-trained on Common Crawl 840B tokens [22]. The maximum length of commands is set to T = 40 and for both visual and linguistic features, channel dimension C = 512 …","url":["https://web2py.iiit.ac.in/research_centres/publications/download/inproceedings.pdf.8b947b8b2836d208.49524f5332315f313135365f46492e706466.pdf"]}
{"year":"2021","title":"H2oloo at TREC 2020: When all you got is a hammer... Deep Learning, Health Misinformation, and Precision Medicine","authors":["R Pradeep, X Ma, X Zhang, H Cui, R Xu, R Nogueira… - Corpus"],"snippet":"… ranking. We adopt all the default settings. We index all the news articles found in the CommonCrawl News crawl from January 1st, 2020 to April 30th, 2020. At inference time, we retrieve the top-k0 (=1000) documents per query …","url":["https://trec.nist.gov/pubs/trec29/papers/h2oloo.DL.HM.PM.pdf"]}
{"year":"2021","title":"HAHA at EmoEvalEs 2021: Sentiment Analysis in Spanish Tweets with Cross-lingual Model","authors":["K Li - Proceedings of the Iberian Languages Evaluation …, 2021"],"snippet":"… It is a large multi-lingual language model, trained on 2.5TB of filtered Common Crawl data … Because the XLM-RoBERTa model increases the number of languages and the number of training data sets, it uses more …","url":["http://ceur-ws.org/Vol-2943/emoeval_paper5.pdf"]}
{"year":"2021","title":"Harmonized System Code Classification Using Transfer Learning with Pre-Trained Weights","authors":["K Pain - 2021"],"snippet":"Page 1. HARMONIZED SYSTEM CODE CLASSIFICATION USING TRANSFER LEARNING WITH PRE-TRAINED WEIGHTS by Koustav Pain (Tukai) Submitted in partial fulfillment of the requirements for the degree of Master of Computer Science at …","url":["https://dalspace.library.dal.ca/bitstream/handle/10222/80672/KoustavPain2021.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Hate Speech Detection based on Sentiment Knowledge Sharing","authors":["X Zhou, Y Yang, X Fan, G Ren, Y Song, Y Diao, L Yang…"],"snippet":"… To compare with previous works, We report results of DV using the standard Accuracy and weighted F1. In our experiments, for the input layer, all word vectors are initialized by Glove Common Crawl Embeddings (840B Token), and the dimension is 300 …","url":["https://aclanthology.org/2021.acl-long.556.pdf"]}
{"year":"2021","title":"HDRS: Hindi Dialogue Restaurant Search Corpus for Dialogue State Tracking in Task-Oriented Environment","authors":["S Malviya, R Mishra, SK Barnwal, US Tiwary - IEEE/ACM Transactions on Audio …, 2021"],"snippet":"Page 1. 2329-9290 (c) 2021 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This …","url":["https://ieeexplore.ieee.org/abstract/document/9376978/"]}
{"year":"2021","title":"HerBERT: Efficiently Pretrained Transformer-based Language Model for Polish","authors":["R Mroczkowski, P Rybak, A Wróblewska, I Gawlik - … of the 8th Workshop on Balto …, 2021"],"snippet":"… CCNet (Wenzek et al., 2020) is a clean monolingual corpus extracted from Common Crawl6 dataset of crawled websites … The first one did not use pretrained embeddings while the latter utilized fastText (Bo- janowski …","url":["https://www.aclweb.org/anthology/2021.bsnlp-1.1.pdf"]}
{"year":"2021","title":"Hierarchical Ranking for Answer Selection","authors":["H Gao, M Hu, R Cheng, T Gao - arXiv preprint arXiv:2102.00677, 2021"],"snippet":"Page 1. Hierarchical Ranking for Answer Selection Hang Gao, Mengting Hu, Renhong Cheng, Tiegang Gao Nankai University knimet@mail.nankai. edu.cn Abstract Answer selection is a task to choose the positive answers …","url":["https://arxiv.org/pdf/2102.00677"]}
{"year":"2021","title":"HIGH PERFORMANCE DOCUMENT STORE IMPLEMENTATION IN RUST","authors":["I Aggarwal - 2021"],"snippet":"Databases are a core part of any application which requires persistence of data. The performance of applications involving the use of database systems is directly proportional to how fast their database read-write operations are. The aim of this …","url":["https://scholarworks.sjsu.edu/cgi/viewcontent.cgi?article=2044&context=etd_projects"]}
{"year":"2021","title":"HISTOGRAM OF TASK UNCERTAINTY","authors":["QL SCHEDULER, MT LEARNING"],"snippet":"… 2014) word embedding (300d version on 840B Common Crawl data), and Xavier initialization for the parameters. The mini-batch size is set to 16, including samples of the same task. Other than these, we follow the training procedure for (Søgaard …","url":["https://openreview.net/pdf?id=sHUFhv03qX_"]}
{"year":"2021","title":"HmBlogs: A big general Persian corpus","authors":["HM Khansari, M Shamsfard - arXiv preprint arXiv:2111.02362, 2021"],"snippet":"… The main part of PRT is obtained from the Persian section of the Common Crawl project (Common Crawl, nd). Common Crawl is a web crawl project that crawls and collects resources available on the web in any language, including some in Persian …","url":["https://arxiv.org/pdf/2111.02362"]}
{"year":"2021","title":"HONEST: Measuring hurtful sentence completion in language models","authors":["D Nozza, F Bianchi, D Hovy - Proceedings of the 2021 Conference of the North …, 2021"],"snippet":"Page 1. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2398–2406 June 6–11, 2021. ©2021 Association for Computational Linguistics 2398 …","url":["https://www.aclweb.org/anthology/2021.naacl-main.191.pdf"]}
{"year":"2021","title":"Hope Speech detection in under-resourced Kannada language","authors":["A Hande, R Priyadharshini, A Sampath, KP Thamburaj… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Advances in Computational Intelligence manuscript No. (will be inserted by the editor) Hope Speech detection in under-resourced Kannada language Adeep Hande · Ruba Priyadharshini · Anbukkarasi Sampath · Kingston …","url":["https://arxiv.org/pdf/2108.04616"]}
{"year":"2021","title":"How do blind people know that blue is cold? Distributional semantics encode color-adjective associations","authors":["J van Paridon, Q Liu, G Lupyan - PsyArXiv. February, 2021"],"snippet":"… trained on. Using the projection method, we computed color associations in embeddings trained on four different text corpora: The Common Crawl and Wikipedia (Grave, Bojanowski, Gupta, Joulin, & Page 2. Mikolov, 2018 …","url":["https://psyarxiv.com/vyxpq/download/?format=pdf"]}
{"year":"2021","title":"How Do Simple Transformations of Text and Image Features Impact Cosine-Based Semantic Match?","authors":["G Collell, MF Moens - ECIR (1), 2021"],"snippet":"… 1. Word-level features: – GloVe6 [40]: We use 300-d vectors pre-trained on the Common Crawl corpus with 840B tokens and a 2.2M-word vocabulary. – word2vec (w2v) [37]: We use the skip-gram 300-d embeddings trained on Wikipedia …","url":["https://www.researchgate.net/profile/Guillem-Collell/publication/350427435_How_Do_Simple_Transformations_of_Text_and_Image_Features_Impact_Cosine-Based_Semantic_Match/links/60851cdd907dcf667bc0b6a8/How-Do-Simple-Transformations-of-Text-and-Image-Features-Impact-Cosine-Based-Semantic-Match.pdf"]}
{"year":"2021","title":"How far can we get with one GPU in 100 hours? CoAStaL at MultiIndicMT Shared Task","authors":["R Aralikatte, HRM Bello, M de Lhoneux, D Hershcovich…"],"snippet":"… translation task. It is pre-trained on publicly available corpora from Wikipedia and Common Crawl. It … tion task. This model is pre-trained on mC4, a multilingual version of the Common Crawl consisting of text from 101 languages. It …","url":["https://aclanthology.org/2021.wat-1.24.pdf"]}
{"year":"2021","title":"How good BERT based models are in sentiment analysis of Croatian tweets: comparison of four multilingual BERTs","authors":["M Ptiček"],"snippet":"… It is pre-trained on texts on 100 different languages and more than two terabytes of data filtered from CommonCrawl data. As stated by the authors “XLM-RoBERTa (XLM-R) outperforms mBERT on cross-lingual classification by up to 23% accuracy on low-resource …","url":["http://archive.ceciis.foi.hr/app/public/conferences/2021/Proceedings/IS/IS1.pdf"]}
{"year":"2021","title":"How Great is the Great Firewall? Measuring China's DNS Censorship","authors":["NP Hoang, AA Niaki, J Dalek, J Knockel, P Lin… - arXiv preprint arXiv …, 2021"],"snippet":"… nsarchive.gwu.edu, cs.colorado.edu) are censored but their SLDs (eg, mit.edu, gwu.edu, colorado.edu) are not. We complement our test list by including domains from the Citizen Lab test lists (CLTL) [13], the Tranco list [66], and the Common Crawl project [14] …","url":["https://arxiv.org/pdf/2106.02167"]}
{"year":"2021","title":"How much is a cow like a meow? A novel database of human judgements of","authors":["K Wegner-Clemens, GL Malcolm, S Shomstein - Psychological Science"],"snippet":"… This model was trained on a total of 650 billion words including Wikipedia from June 2017, two news corpuses (statmt.org news, UMBC news), and corpuses derived from a wide range of websites (Gigagword, Common Crawl). The words …","url":["https://psyarxiv.com/7h82c/download/?format=pdf"]}
{"year":"2021","title":"How to Adapt Your Pretrained Multilingual Model to 1600 Languages","authors":["A Ebrahimi, K Kann - arXiv preprint arXiv:2106.02124, 2021"],"snippet":"… 2018) who present embeddings for 157 languages trained on additional data from Common Crawl … the larger vocabulary of 250k subwords created using SentencePiece tokenization (Kudo and Richardson, 2018) and the …","url":["https://arxiv.org/pdf/2106.02124"]}
{"year":"2021","title":"How Viable is Password Cracking in Digital Forensic Investigation? Analyzing the Guessability of over 3.9 Billion Real-World Accounts","authors":["A Kanta, S Coray, I Coisel, M Scanlon - Forensic Science International: Digital …, 2021"],"snippet":"… GloVe [34] was used to automate this process with its Common Crawl 42B 300d, a pre-trained model in English for GloVe15 … This is because they do not have a representation in the Common Crawl and therefore cannot be compared to the categories …","url":["https://www.forensicsandsecurity.com/papers/PasswordCracking3BillionAccounts.pdf"]}
{"year":"2021","title":"How Vulnerable Are Automatic Fake News Detection Methods to Adversarial Attacks?","authors":["C Koenders, J Filla, N Schneider, V Woloszyn - arXiv preprint arXiv:2107.07970, 2021"],"snippet":"Page 1. How Vulnerable Are Automatic Fake News Detection Methods to Adversarial Attacks? Camille Koenders Technische Universität Berlin camille.koenders@gmail.com Johannes Filla Technische Universität Berlin johannesfilla@gmx.de …","url":["https://arxiv.org/pdf/2107.07970"]}
{"year":"2021","title":"HowSumm: A Multi-Document Summarization Dataset Derived from WikiHow Articles","authors":["O Boni, G Feigenblat, G Lev, M Shmueli-Scheuer… - arXiv preprint arXiv …, 2021"],"snippet":"… a corpus. They es- tablish this approach by extracting answers that are Wikipedia paragraphs from Google Natural Questions (Kwiatkowski et al., 2019) and documents from Common Crawl (Raffel et al., 2020) corpus. In some …","url":["https://arxiv.org/pdf/2110.03179"]}
{"year":"2021","title":"HTLM: Hyper-Text Pre-Training and Prompting of Language Models","authors":["A Aghajanyan, D Okhonko, M Lewis, M Joshi, H Xu… - arXiv preprint arXiv …, 2021"],"snippet":"… Hyper-text, such as the HTML found in the Common Crawl1, has a number of advantages for pretraining over plain text … Our HyperText Language Model (HTLM) is trained on 23TB of simplified HTML which we automatically extract …","url":["https://arxiv.org/pdf/2107.06955"]}
{"year":"2021","title":"HULTIG-C: NLP Corpus and Services in the Cloud","authors":["S Pais, J Cordeiro, M Jamil - 2021"],"snippet":"… [11] presented a Multilingual Web Corpus in over 50 languages with over 10 billion tokens licensed under the Creative Commons license. The texts that made up the corpus were extracted from Common Crawl, with around 2 billion crawled URLs …","url":["https://www.researchsquare.com/article/rs-696114/latest.pdf"]}
{"year":"2021","title":"HumAID: Human-Annotated Disaster Incidents Data from Twitter","authors":["F Alam, U Qazi, M Imran, F Ofli - arXiv preprint arXiv:2104.03090, 2021"],"snippet":"Page 1. HumAID: Human-Annotated Disaster Incidents Data from Twitter with Deep Learning Benchmarks Firoj Alam, Umair Qazi, Muhammad Imran, Ferda Ofli Qatar Computing Research Institute, HBKU, Qatar {fialam,uqazi,mimran,fofli}@hbku.edu.qa …","url":["https://arxiv.org/pdf/2104.03090"]}
{"year":"2021","title":"Human and Transformer-Based Prosodic Phrasing in Two Speech Genres","authors":["J Volín, M Řezáčková, J Matouřek - International Conference on Speech and …, 2021"],"snippet":"… We worked with the pre-trained Text-to-Text TransferTransformer (T5) model [18] and used the ability of weights pre-training from a huge amount of unlabeled text in the language [4, 18] – for this task we used Czech …","url":["https://link.springer.com/chapter/10.1007/978-3-030-87802-3_68"]}
{"year":"2021","title":"Human-Centered Financial Summarization","authors":["T PASSALI - 2021"],"snippet":"Page 1. ARISTOTLE UNIVERSITY OF THESSALONIKI MASTER'S THESIS Human-Centered Financial Summarization Author: Tatiana PASSALI Supervisor: Dr. Grigorios TSOUMAKAS A thesis submitted in fulfillment of …","url":["http://ikee.lib.auth.gr/record/329076/files/GRI-2021-30128.pdf"]}
{"year":"2021","title":"Hybrid approach to detecting symptoms of depression in social media entries","authors":["A Wołk, K Chlasta, P Holas - arXiv preprint arXiv:2106.10485, 2021"],"snippet":"… 2004). To conduct the experiments, we trained our model on the data from the Common Crawl (Bevendorff et al. 2018) project. As the CommonCrawl corpus is very polluted, an algorithmic division into sentences, data de-duplication, and cleaning was necessary …","url":["https://arxiv.org/pdf/2106.10485"]}
{"year":"2021","title":"I Wish I Would Have Loved This One, But I Didn't--A Multilingual Dataset for Counterfactual Detection in Product Reviews","authors":["J O'Neill, P Rozenshtein, R Kiryo, M Kubota… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. arXiv:2104.06893v1 [cs.CL] 14 Apr 2021 I Wish I Would Have Loved This One, But I Didn't – A Multilingual Dataset for Counterfactual Detection in Product Reviews James O'Neill‡,∗∗Polina Rozenshtein†,∗ Ryuichi …","url":["https://arxiv.org/pdf/2104.06893"]}
{"year":"2021","title":"IBM MNLP IE at CASE 2021 task 1: Multi-granular and multilingual event detection on protest news","authors":["P Awasthy, J Ni, K Barker, R Florian - Proceedings of the 4th Workshop on …, 2021"],"snippet":"… Hence we decided to use XLM-R as the text encoder. We use HuggingFace's pytorch im- plementation of transformers (Wolf et al., 2019). XLM-R was pre-trained with unlabeled Wikipedia text and the CommonCrawl Corpus of 100 languages …","url":["https://aclanthology.org/2021.case-1.18.pdf"]}
{"year":"2021","title":"iCompass at NLP4IF-2021–Fighting the COVID-19 Infodemic","authors":["W Henia, O Rjab, H Haddad, C Fourati - Proceedings of the Fourth Workshop on NLP …, 2021"],"snippet":"… Page 3. 117 3.4 Arabic ALBERT Arabic ALBERT2 by (KUIS-AI-Lab) models were pretrained on 4.4 Billion words: Arabic version of OSCAR (unshuffled version of the corpus) filtered from Common Crawl and Recent dump of Arabic Wikipedia …","url":["https://www.aclweb.org/anthology/2021.nlp4if-1.17.pdf"]}
{"year":"2021","title":"Identification of Biased Terms in News Articles by Comparison of Outlet-specific Word Embeddings","authors":["T Spinde, L Rudnitckaia, F Hamborg"],"snippet":"… We scraped articles from both news outlets, published in the last decade, from 2010 to 2020, from Common Crawl [4]. For preprocessing, we use Genism simple preprocessing and generate n-grams. 3.4 Linear mapping between vector spaces …","url":["https://www.researchgate.net/profile/Timo_Spinde2/publication/348303058_Identification_of_Biased_Terms_in_News_Articles_by_Comparison_of_Outlet-specific_Word_Embeddings/links/5ff70425a6fdccdcb837e519/Identification-of-Biased-Terms-in-News-Articles-by-Comparison-of-Outlet-specific-Word-Embeddings.pdf"]}
{"year":"2021","title":"Identification of Toxic, Engaging, and Fact-Claiming Comments","authors":["J Risch, A Stoll, L Wilms, M Wiegand - 2021"],"snippet":"Page 1. Proceedings of the GermEval 2021 Workshop on the Identification of Toxic, Engaging, and Fact-Claiming Comments 17th Conference on Natural Language Processing KONVENS 2021 Heinrich Heine University Düsseldorf …","url":["https://netlibrary.aau.at/obvukloa/content/titleinfo/6435190/full.pdf"]}
{"year":"2021","title":"Identifying communicative functions in discourse with content types","authors":["T Caselli, R Sprugnoli, G Moretti - Language Resources and Evaluation, 2021"],"snippet":"Texts are not monolithic entities but rather coherent collections of micro illocutionary acts which help to convey a unitary message of content and purpose.","url":["https://link.springer.com/article/10.1007/s10579-021-09554-4"]}
{"year":"2021","title":"Idiomatic Expression Identification using Semantic Compatibility","authors":["Z Zeng, S Bhat - arXiv preprint arXiv:2110.10064, 2021"],"snippet":"Idiomatic expressions are an integral part of natural language and constantly being added to a language. Owing to their non-compositionality and their ability to take on a figurative or literal meaning depending on the sentential context, they have been a …","url":["https://arxiv.org/pdf/2110.10064"]}
{"year":"2021","title":"IIITT@ DravidianLangTech-EACL2021: Transfer Learning for Offensive Language Detection in Dravidian Languages","authors":["K Yasaswini, K Puranik, A Hande, R Priyadharshini… - Proceedings of the First …, 2021"],"snippet":"… 4.3 XLM-RoBERTa XLM-RoBERTa (Conneau et al., 2020) is a large multi-lingual language model, trained on 2.5TB of cleaned CommonCrawl data in 100 languages. It can be recognized as a union of XLM (Lample …","url":["https://www.aclweb.org/anthology/2021.dravidianlangtech-1.25.pdf"]}
{"year":"2021","title":"IIITT@ LT-EDI-EACL2021-Hope Speech Detection: There is always hope in Transformers","authors":["K Puranik, A Hande, R Priyadharshini, S Thavareesan… - arXiv preprint arXiv …, 2021"],"snippet":"… It's trained on over 2TB of filtered CommonCrawl data in 100 different languages. It was an update to the XLM-100 model (Lample and Conneau, 2019) but with in- creased training data … Wikipedia, Common Crawl7, PMINDIA8 and Dakshina9 datasets …","url":["https://arxiv.org/pdf/2104.09066"]}
{"year":"2021","title":"Image Emotion Recognition Using Visual and Semantic Features Reflecting Emotional and Similar Objects","authors":["T YAMAMOTO, S TAKEUCHI, A NAKAZAWA - IEICE TRANSACTIONS on Information …, 2021"],"snippet":"… GloVe has been used for sentiment analysis [40]; therefore, we use GloVe as the method for obtaining the word vector. It has a 300-dimensional word embedding weight pre-trained by Common Crawl (840 billion tokens), which is publicly available [41] …","url":["https://www.jstage.jst.go.jp/article/transinf/E104.D/10/E104.D_2020EDP7218/_pdf"]}
{"year":"2021","title":"Impacts of Homophone Normalization on Semantic Models for Amharic","authors":["TD Belay, E Kombolcha, AA Ayele, G Gelaye…"],"snippet":"… embedding representation. The only available pretrained static word embedding for Amharic text is the fastText model, which is trained from Wikipedia and web data scraped in the common crawl project [19]. This embedding …","url":["https://www.inf.uni-hamburg.de/en/inst/ab/lt/publications/2021-belayetal-ict4da-amharicnorm.pdf"]}
{"year":"2021","title":"Implicit social cognition: A brief (and gentle) introduction","authors":["B Kurdi, MR Banaji - PsyArXiv. January, 2021"],"snippet":"… individual-level implicit attitudes toward ten social categories indexed using the IAT are closely aligned with societal-level semantic associations derived from the vast Common Crawl corpus via so- called word embeddings …","url":["https://psyarxiv.com/a4pjy/download/?format=pdf"]}
{"year":"2021","title":"Implicitly Abusive Comparisons–A New Dataset and Linguistic Analysis","authors":["M Wiegand, M Geulig, J Ruppenhofer - Proceedings of the 16th Conference of the …, 2021"],"snippet":"… A comparison was represented by the embedding vector of the word in that comparison whose similarity was highest to the centroid. As embeddings we chose the fastText emeddings (Joulin et al., 2017) induced on Common Crawl.8 Emotions (EMO) …","url":["https://www.aclweb.org/anthology/2021.eacl-main.27.pdf"]}
{"year":"2021","title":"Improved Automated Classification of Sentences in Data Science Exercises","authors":["AM Angelone, A Galassi, P Vittorini - … in Methodologies and intelligent Systems for …, 2021"],"snippet":"… In our setting, fastText embeddings are vectors of 300 dimensions, BERT embeddings are vectors of 768 dimensions. As for fastText, we adopted the precomputed Italian language model 1 , trained on Common Crawl and Wikipedia …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86618-1_2"]}
{"year":"2021","title":"Improved Word Representations Via Summed Target and Context Embeddings","authors":["N Fulda, N Robinson"],"snippet":"… The word2vec and word2vec-PLUS scores here were trained on the corpus of scraped ar- ticles since it resembles the Common Crawl data used to train GloVe [9] more closely than the other, more specialized corpora we explored …","url":["https://dragn.ai/wp-content/uploads/2021/01/sami2021_39_Nancy-Fulda.pdf"]}
{"year":"2021","title":"Improving Hierarchical Product Classification using Domain-specific Language Modelling","authors":["A Brinkmann, C Bizer - 2021"],"snippet":"… Afterwards, we investigate whether it is possible to improve the performance of the transformer models by performing additional self-supervised pre-training using different corpora of product offers, which were extracted from the Common Crawl …","url":["https://www.uni-mannheim.de/media/Einrichtungen/dws/Files_Research/Web-based_Systems/pub/Brinkmann-Bizer-Improving_Hierarchical_Product_Classification_using_domain_specific_laguage_modelling-PKG4Ecommerce2021.pdf"]}
{"year":"2021","title":"Improving Large-scale Language Models and Resources for Filipino","authors":["JCB Cruz, C Cheng - arXiv preprint arXiv:2111.06053, 2021"],"snippet":"… 2019) is a massive dataset obtained from language identification and filtering of the Common Crawl dataset. We use the deduplicated version of the Filipino (Tagalog) portion of OSCAR and add it to our pretraining corpus. …","url":["https://arxiv.org/pdf/2111.06053"]}
{"year":"2021","title":"Improving Multilingual Models for the Swedish Language: Exploring CrossLingual Transferability and Stereotypical Biases","authors":["S Katsarou - 2021"],"snippet":"… To acquire the natural text needed for pretraining purposes, the authors used common crawl [21]. Common crawl is a non commercial … To justify the choice of Common Crawl the authors say that it is a publicly available repository, allowing any …","url":["https://www.diva-portal.org/smash/get/diva2:1618310/FULLTEXT01.pdf"]}
{"year":"2021","title":"Improving question answering for event-focused questions in temporal collections of news articles","authors":["J Wang, A Jatowt, M Färber, M Yoshikawa - Information Retrieval Journal"],"snippet":"Temporal collections of news articles (or news archives) contain numerous accurate and time-aligned articles, which offer immense value to our society, hel.","url":["https://link.springer.com/article/10.1007/s10791-020-09387-9"]}
{"year":"2021","title":"Improving Sentiment Classification for Arabic Short Text Using Deep Learning Approaches","authors":["A Alwehaibi - 2021"],"snippet":"Page 1. Improving Sentiment Classification for Arabic Short Text Using Deep Learning Approaches Ali Alwehaibi North Carolina A&T State University A dissertation submitted to the graduate faculty in partial fulfillment of the requirements for the degree of …","url":["https://search.proquest.com/openview/615ff8051266a66aec4a923d0faafa78/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Improving the Lexical Ability of Pretrained Language Models for Unsupervised Neural Machine Translation","authors":["A Chronopoulou, D Stojanovski, A Fraser - arXiv preprint arXiv:2103.10531, 2021"],"snippet":"… We use 3K randomly sampled sentences of SETIMES (Tiedemann, 2012) as validation/test sets. We also use 68M En sentences from NewsCrawl. For Sq and Mk we use all the CommonCrawl corpora from Ortiz Suárez et …","url":["https://arxiv.org/pdf/2103.10531"]}
{"year":"2021","title":"in the Language Technology (LT) group","authors":["T Fischer, C Biemann, SM Yimam"],"snippet":"… The model is trained on the colossal and cleaned version of common crawl (C4) as well as HugeNews which consists of 1.5B news articles from the web. Unsurprisingly, this method achieves state-of-the-art results on many summarization datasets …","url":["https://www.inf.uni-hamburg.de/en/inst/ab/lt/teaching/theses/completed-theses/2021-ma-timfischer.pdf"]}
{"year":"2021","title":"indicnlp@ kgp at DravidianLangTech-EACL2021: Offensive Language Identification in Dravidian Languages","authors":["K Kedia, A Nandy - arXiv preprint arXiv:2102.07150, 2021"],"snippet":"… use the bert-base-multilingual-cased model, mBERT trained on cased text in 104 languages from large Wikipedia articles, and the xlm-roberta-base model, XLM-R (Conneau et al., 2020) trained on 100 languages, using …","url":["https://arxiv.org/pdf/2102.07150"]}
{"year":"2021","title":"IndoNLG: Benchmark and Resources for Evaluating Indonesian Natural Language Generation","authors":["S Cahyawijaya, GI Winata, B Wilie, K Vincentio, X Li… - arXiv preprint arXiv …, 2021"],"snippet":"… Wiki Javanese1 6,015,961 231,571 53.2 MB formal Wikipedia CC-100 Sundanese 13,761,754 433,086 107.6 MB mixed Common Crawl CC-100 Javanese 20,560,458 690,517 161.9 MB mixed Common Crawl TOTAL 3,626,283,931 276,838,931 23.79 GB …","url":["https://arxiv.org/pdf/2104.08200"]}
{"year":"2021","title":"Information Extraction from Semi-Structured Websites","authors":["C Lockard - 2021"],"snippet":"… Note the log scale of the x-axis. . . . . 42 3.4 Extraction precision vs. number of extractions on the CommonCrawl dataset at various confidence thresholds; the orange bar indicates a 0.75 threshold yielding 1.25millionextractionsat90%precision. . . . . 47 …","url":["https://digital.lib.washington.edu/researchworks/bitstream/handle/1773/47422/Lockard_washington_0250E_22994.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Information Retrieval Model for Social Media Applications","authors":["S Bhandari - 2021"],"snippet":"… 2019 IEEE International Conference on Healthcare Informatics (ICHI), pp. 1–3, 2019. [17] S. Nagel, “Cc-news.” https://commoncrawl.org/2016/10/news-datasetavailable/, 2016. [18] M. Heidari and JH Jones, “Using bert to extract …","url":["http://129.174.21.13/bitstream/handle/1920/11956/bhandari_social-media-ml.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Information retrieval versus deep learning approaches for generating traceability links in bilingual projects","authors":["J Lin, Y Liu, J Cleland-Huang - Empirical Software Engineering, 2022"],"snippet":"Abstract Software traceability links are established between diverse artifacts of the software development process in order to support tasks such as compliance analysis, safety assurance, and requirements validation. However, practice has shown that it …","url":["https://link.springer.com/article/10.1007/s10664-021-10050-0"]}
{"year":"2021","title":"INFOTEC-LaBD at PAN@ CLEF21: Profiling Hate Speech Spreaders on Twitter through Emotion-based Representations","authors":["H Cabrera, S Miranda-Jiménez, ES Tellez - CLEF, 2021"],"snippet":"… 26] for mapping Tweets to semantic space. In particular, we use the model pre-trained on 600 billion tokens on Common Crawl for the English task and the one from [27] to tackle the Spanish task. We use the NRC Lexicon [12 …","url":["http://ceur-ws.org/Vol-2936/paper-159.pdf"]}
{"year":"2021","title":"Infusing Future Information into Monotonic Attention Through Language Models","authors":["MA Zaidi, S Indurthi, B Lee, NK Lakumarapu, S Kim - arXiv preprint arXiv:2109.03121, 2021"],"snippet":"Page 1. Infusing Future Information into Monotonic Attention Through Language Models Mohd Abbas Zaidi∗, Sathish Indurthi∗†, Beomseok Lee Nikhil Kumar Lakumarapu, Sangha Kim NLP Lab, Samsung Research, Seoul, South …","url":["https://arxiv.org/pdf/2109.03121"]}
{"year":"2021","title":"Input Augmentation Improves Constrained Beam Search for Neural Machine Translation: NTT at WAT 2021","authors":["K Chousa, M Morishita - arXiv preprint arXiv:2106.05450, 2021"],"snippet":"… We also used CommonCrawl provided by the WMT 2020 news shared task (Barrault et al., 2020) as additional monolingual data. For CommonCrawl data, we chose the ten million English and Japanese sentences that are similar …","url":["https://arxiv.org/pdf/2106.05450"]}
{"year":"2021","title":"Integrated Training for Sequence-to-Sequence Models Using Non-Autoregressive Transformer","authors":["E Tokarchuk, J Rosendahl, W Wang, P Petrushkov…"],"snippet":"… Figure 2: Different variants of the encoder-decoder model integration through the connection interface. 5.1 Data Training data for French→German includes Eu- roparl corpus version 7 (Koehn, 2005), CommonCrawl2 corpus and the newstest2008-2010 …","url":["https://www-i6.informatik.rwth-aachen.de/publications/download/1194/TokarchukEvgeniiaRosendahlJanWangWeiyuePetrushkovPavelLancewickiTomerKhadiviShahramNeyHermann--IntegratedTrainingforSequence-to-SequenceModelsUsingNon-AutoregressiveTransformer--2021.pdf"]}
{"year":"2021","title":"Intelligent Arabic-Based Healthcare Assistant","authors":["T Wael, A Hesham, M Youssef, O Adel, H Hesham… - 2021 3rd Novel Intelligent …, 2021"],"snippet":"… It was pre-trained on ∼8.2 billion words from the Arabic version of OSCAR (unshuffled version of the corpus) - filtered from Common Crawl [6], the recent dump of Arabic Wikipedia, and other Arabic resources, which sum up to ∼95GB of text. The corpus …","url":["https://ieeexplore.ieee.org/abstract/document/9600526/"]}
{"year":"2021","title":"Intelligent Document Validation","authors":["E de Souza Pais - 2021"],"snippet":"Abstract Processes in organizations over the past few years have been increasingly automated. In order to make them more efficient and practical. However, one area in which manual work is still common is document analysis. In this area, due to the …","url":["https://estudogeral.uc.pt/retrieve/217107/Eduardo%20de%20Sousa%20Pais.pdf"]}
{"year":"2021","title":"Intelligent Methods for Accurately Detecting Phishing Websites","authors":["M Al-kasassbeh, ABUZ AL MAHA"],"snippet":"… A. Data set The data set used in this paper is offered by [19] is composed of 5000 phishing websites and 5000 legitimate websites. Phishing websites are collected from PhishTank and OpenPhish, while legitimate websites are collected from Alexa and Common Crawl …","url":["https://www.academia.edu/download/65538093/200207223.pdf"]}
{"year":"2021","title":"Interactive Clustering of Cooking Recipe Instructions: Towards the Automatic Detection of Events Involving Kitchen Devices","authors":["F Ventirozos, M Jacobo-Romero, S Clinch… - 2021 IEEE 15th …, 2021"],"snippet":"… Based on these synonym sets, we enriched our list by adding similar words based on GloVe word embedding vectors [17], which were pre-trained on the Common Crawl corpus1 … 1https://commoncrawl.org/ 342 Page 3. Figure …","url":["https://ieeexplore.ieee.org/abstract/document/9364446/"]}
{"year":"2021","title":"Interactive query expansion for professional search applications","authors":["T Russell-Rose, P Gooch, U Kruschwitz - arXiv preprint arXiv:2106.13528, 2021"],"snippet":"… The DBpedia data set describes 4.58 million entities, out of which 4.22 million are classified in a consistent ontology. 2. WebISA (Seitner et al., 2016) is a publicly available database containing hypernymy relations extracted from the CommonCrawl web corpus …","url":["https://arxiv.org/pdf/2106.13528"]}
{"year":"2021","title":"Internet-Augmented Dialogue Generation","authors":["M Komeili, K Shuster, J Weston - arXiv preprint arXiv:2107.07566, 2021"],"snippet":"… First, we store and utilize the Common Crawl dump of the internet from Wenzek et al … DPR+FAISS-based models We trained DPR+FAISS-based models using either the WoW or WizInt training datasets, and using either …","url":["https://arxiv.org/pdf/2107.07566"]}
{"year":"2021","title":"Interpretability in Word Sense Disambiguation using Tsetlin Machine","authors":["RK Yadav, L Jiao, OC Granmo, M Goodwin - 2021"],"snippet":"… FastText-Base (FTX- B), FastText-CommonCrawl (FTX-C), 1 neural network (NN) BERT base, and our proposed TM. FTX- B is a fast text linear classifier without pre-trained embeddings and FTX-C is a fast text linear classifier with …","url":["https://www.scitepress.org/Papers/2021/103821/103821.pdf"]}
{"year":"2021","title":"Interpretable bias mitigation for textual data: Reducing gender bias in patient notes while maintaining classification performance","authors":["JR Minot, N Cheney, M Maier, DC Elbers, CM Danforth… - arXiv preprint arXiv …, 2021"],"snippet":"… and resulting word embeddings. For instance, general purpose language models are often trained on Wikipedia and the Common Crawl collection of web pages (eg, BERT [27], RoBERTa [28]). Training language models on …","url":["https://arxiv.org/pdf/2103.05841"]}
{"year":"2021","title":"Interpreting and Controlling Linguistic Features in Neural Networks' Representations","authors":["T Limisiewicz"],"snippet":"Abstract In recent years, Neural Networks proved their usefulness in computing and processing vector representations to solve various NLP tasks. The key objective of the thesis is to provide a better explanation of the neural network’s representation of …","url":["https://ufal.mff.cuni.cz/~zabokrtsky/pgs/thesis_proposal/tomasz-limisiewicz-proposal.pdf"]}
{"year":"2021","title":"Introducing A large Tunisian Arabizi Dialectal Dataset for Sentiment Analysis","authors":["C Fourati, H Haddad, A Messaoudi, M BenHajhmida… - Proceedings of the Sixth …, 2021"],"snippet":"… In this pa- per, we introduce a large common-crawl-based Tunisian Arabizi dialectal dataset dedicated for Sentiment Analysis … Hence, our contributions can be summarized as: the first re- lease of a Tunizi large-scale Common-Crawl-based Page 3. 228 …","url":["https://www.aclweb.org/anthology/2021.wanlp-1.25.pdf"]}
{"year":"2021","title":"Introducing huBERT","authors":["ND Márk"],"snippet":"… Abstract This paper introduces the huBERT family of models. The flagship is the eponymous BERT Base model trained on the new Hungarian Webcorpus 2.0, a 9-billion-token corpus of Web text collected from the Common Crawl …","url":["https://hlt.bme.hu/media/pdf/huBERT.pdf"]}
{"year":"2021","title":"Introduction: digital humanities and the use of web archives","authors":["K Teszelszky - 2021"],"snippet":"… Also, an ever-increasing amount of data is selected, archived, and presented worldwide in web collections by organizations like the Internet Archive, Common Crawl, national libraries, and other academic research and digital heritage …","url":["https://link.springer.com/article/10.1007/s42803-021-00040-5"]}
{"year":"2021","title":"Investigating Depression Semantics on Reddit","authors":["S Agarwal, P Singh, J Shah, N Sanjeev - International Conference on Neural …, 2021"],"snippet":"Major depression is a challenging issue affecting individuals and those of the people around them. This paper investigates the Reddit comments for the automated identification of comments being indicative of depressive behaviour. We …","url":["https://link.springer.com/chapter/10.1007/978-3-030-92310-5_75"]}
{"year":"2021","title":"Investigating Hostile Post Detection in Hindi","authors":["V Bhatnagar, P Kumar, P Bhattacharyya - Neurocomputing, 2021"],"snippet":"… Wiki + CC : Pretrained Embeddings trained on Wikipedia and Common Crawl Wiki : Pretrained Embeddings trained on Wikipedia TEAP : … Wikipedia pretrained word vectors 25 and Wikipedia + Common Crawl pretrained word vectors 26 for …","url":["https://www.sciencedirect.com/science/article/pii/S0925231221018099"]}
{"year":"2021","title":"Investigating Monolingual and Multilingual BERTModels for Vietnamese Aspect Category Detection","authors":["D Van Thin, LS Le, VX Hoang, NLT Nguyen - arXiv preprint arXiv:2103.09519, 2021"],"snippet":"… Besides, there are many variants of BERT models which are published for Vietnamese language and multi-lingual model as cross-lingual language model (XLM-R) [10] or multilingual BERT [9]. These models are pre-trained …","url":["https://arxiv.org/pdf/2103.09519"]}
{"year":"2021","title":"Investigation of different decoding algorithms for the automatic generation of alternative language patterns in German to English translation","authors":["M Zanon"],"snippet":"… Crawled Corpus (C4) to pretrain the model on. They base this data set on the text data from April 2019 of the Common Crawl archive, which provides web extracted text by scraping about 20TB of text from HTML files each month …","url":["https://users.informatik.haw-hamburg.de/~schumann/BachelorArbeitMelinaZanon.pdf"]}
{"year":"2021","title":"Is Domain Adaptation Worth Your Investment? Comparing BERT and FinBERT on Financial Tasks","authors":["B Peng, E Chersoni, YY Hsu, CR Huang - Proceedings of the Third Workshop on …, 2021"],"snippet":"… The first-generation Transformers were mainly trained on general corpora, such as Wikipedia or Common Crawl. However, considering domain adaptations, many researchers have later injected domain-specific knowledge in such architectures …","url":["https://aclanthology.org/2021.econlp-1.5.pdf"]}
{"year":"2021","title":"Is the Number of Trainable Parameters All That Actually Matters?","authors":["A Chatelain, A Djeghri, D Hesslow, J Launay, I Poli - arXiv preprint arXiv:2109.11928, 2021"],"snippet":"… Architecture details for the four model sizes considered are available in the supplementary. The model is trained on French data obtained using a Common Crawl dump filtered by CCNet [23]. We use byte-level byte-pair encoding, with a vocabulary of 50,262 tokens …","url":["https://arxiv.org/pdf/2109.11928"]}
{"year":"2021","title":"isiZulu Word Embeddings","authors":["S Dlamini, E Jembere, A Pillay, B van Niekerk - 2021 Conference on Information …, 2021"],"snippet":"… TABLE 1: THE MODELS TO BE COMPARED ALONG WITH DETAILS OF THE CORPORA THEY WERE TRAINED ON. Model Corpus Corpus Size Vocabulary Word2Vec Google News 6 billion 1 million GloVe Common …","url":["https://ieeexplore.ieee.org/abstract/document/9395011/"]}
{"year":"2021","title":"JABER: Junior Arabic BERt","authors":["A Ghaddar, Y Wu, A Rashid, K Bibi, M Rezagholizadeh… - arXiv preprint arXiv …, 2021"],"snippet":"… 2020) is a 12 layer BERT model trained on 95GB of common crawl, news, and Wikipedia Arabic data. AraBERT (Antoun et al.… • Common Crawl (CC) This data was downloaded from 10 shards of monthly Common Crawl 3 covering March to …","url":["https://arxiv.org/pdf/2112.04329"]}
{"year":"2021","title":"Jigsaw: Large Language Models meet Program Synthesis","authors":["N Jain, S Vaidyanath, A Iyer, N Natarajan… - arXiv preprint arXiv …, 2021"],"snippet":"Large pre-trained language models such as GPT-3, Codex, and Google's language model are now capable of generating code from natural language specifications of programmer intent. We view these developments with a mixture of optimism and …","url":["https://arxiv.org/pdf/2112.02969"]}
{"year":"2021","title":"JMIR Mental Health","authors":["A Wongkoblap, M Vadillo, V Curcin, C Rauschenberg… - Mental Health"],"snippet":"… occurrent words as vectors [25]. GloVe was trained on several textual data sets, such as Wikipedia and common crawl (a copy of web content), and supported 50D, 100D, 200D, and 300D vectors. Anaphora resolution is another …","url":["https://mental.jmir.org/2021/8/PDF"]}
{"year":"2021","title":"Joint Open Knowledge Base Canonicalization and Linking","authors":["Y Liu, W Shen, Y Wang, J Wang, Z Yang, X Yuan - Proceedings of the 2021 …, 2021"],"snippet":"… 1 − Simemb (si,sj ) if xij = 0 For instance, the score ofSimemb (“Barack Obama”, “President Obama”) is 0.873 using fastText [17] embeddings trained on Common Crawl via MindSpore Framework1, which indicates these two NPs are likely to refer to the same entity …","url":["https://dl.acm.org/doi/abs/10.1145/3448016.3452776"]}
{"year":"2021","title":"Joint source–target encoding with pervasive attention","authors":["M Elbayad, L Besacier, J Verbeek - Machine Translation, 2021"],"snippet":"The pervasive attention model is a sequence-to-sequence model that addresses the issue of source–target interaction in encoder–decoder models by jointly encoding the two sequences with a two-dimensional convolutional neural network. We …","url":["https://link.springer.com/article/10.1007/s10590-021-09289-7"]}
{"year":"2021","title":"Juris2vec: Building Word Embeddings from Philippine Jurisprudence","authors":["E Peramo, C Cheng, M Cordel - … International Conference on Artificial Intelligence in …, 2021"],"snippet":"… utilized a 2010 Wikipedia dump with 1 billion tokens; a 2014 Wikipedia dump with 1.6 billion tokens; Gigaword 5 which has 4.3 billion tokens; the combination Gigaword5 + Wikipedia2014, which has 6 billion tokens; and …","url":["https://ieeexplore.ieee.org/abstract/document/9415251/"]}
{"year":"2021","title":"Just Say No: Analyzing the Stance of Neural Dialogue Generation in Offensive Contexts","authors":["A Baheti, M Sap, A Ritter, M Riedl - arXiv preprint arXiv:2108.11830, 2021"],"snippet":"… For example, OpenAI's GPT-3 (Brown et al., 2020) is a 175 billion parameter neural network, trained on a 570GB subset CommonCrawl that is capable of engaging in impressive open-domain di- alogues with a user, when …","url":["https://arxiv.org/pdf/2108.11830"]}
{"year":"2021","title":"Just What do You Think You're Doing, Dave?'A Checklist for Responsible Data Use in NLP","authors":["A Rogers, T Baldwin, K Leins - arXiv preprint arXiv:2109.06598, 2021"],"snippet":"Page 1. 'Just What do You Think You're Doing, Dave?' ∗ A Checklist for Responsible Data Use in NLP Anna Rogers Center for Social Data Science University of Copenhagen Denmark arogers@sodas.ku.dk Timothy Baldwin …","url":["https://arxiv.org/pdf/2109.06598"]}
{"year":"2021","title":"KAN: Knowledge-Augmented Networks for Few-Shot Learning","authors":["Z Zhu, X Lin - ICASSP 2021-2021 IEEE International Conference on …, 2021"],"snippet":"… 3.3. Hyperparameters For the knowledge graph, to represent the initial semantic features of nodes, we use the GloVe model [23] which is the Common Crawl version trained on 840B tokens. The dimension of the semantic features is 300 …","url":["https://ieeexplore.ieee.org/abstract/document/9413612/"]}
{"year":"2021","title":"KazNERD: Kazakh Named Entity Recognition Dataset","authors":["R Yeshpanov, Y Khassanov, HA Varol - arXiv preprint arXiv:2111.13419, 2021"],"snippet":"We present the development of a dataset for Kazakh named entity recognition. The dataset was built as there is a clear need for publicly available annotated corpora in Kazakh, as well as annotation guidelines containing straightforward--but rigorous--rules …","url":["https://arxiv.org/pdf/2111.13419"]}
{"year":"2021","title":"KI2TE: Knowledge-Infused InterpreTable Embeddings for COVID-19 Misinformation Detection","authors":["W Shiao, EE Papalexakis"],"snippet":"… After cleaning, we were left with around 20K fake news articles. 3.4.3 Real News Dataset. We used the list from the BS Detector Chrome extension3 to pick the reputable sites. We then collected ar- ticles from all of the matching …","url":["https://www.cs.ucr.edu/~epapalex/papers/Knod2021_paper_7.pdf"]}
{"year":"2021","title":"KIT's IWSLT 2021 Offline Speech Translation System","authors":["TN Nguyen, TS Nguyen, C Huber, M Awiszus… - Proceedings of the 18th …, 2021"],"snippet":"… On the other hand, Page 2. 126 Table 2: Text Training Data Dataset Sentences TED Talks (TED) 220K Europarl (EPPS) 2.2MK CommonCrawl 2.1M Rapid 1.21M ParaCrawl 25.1M OpenSubtitles 12.6M WikiTitle 423K Back-translated News 26M …","url":["https://aclanthology.org/2021.iwslt-1.13.pdf"]}
{"year":"2021","title":"Knowledge Graphs for Analyzing and Searching Legal Data","authors":["K Sabrina, P Axel - 2021"],"snippet":"Page 1. Titel der Dissertation: Knowledge Graphs for Analyzing and Searching Legal Data Dissertation zur Erlangung des akademischen Grades eines Doktors der Sozialund Wirtschaftswissenschaften an der Wirtschaftsuniversität Wien eingereicht bei …","url":["https://aic.ai.wu.ac.at/~polleres/supervised_theses/Erwin_Filtz_PhD_2021.pdf"]}
{"year":"2021","title":"Knowledge-driven Answer Generation for Conversational Search","authors":["M Leite, R Ferreira, D Semedo, J Magalhães - arXiv preprint arXiv:2104.06892, 2021"],"snippet":"… This model is pre-trained on the large C4 cor- 5 Page 6. pus, which was derived from Common Crawl1. A masked language modelling objective is used, where the model is trained to predict corrupted randomly sampled tokens of varying sizes. 5. Evaluation 5.1 …","url":["https://arxiv.org/pdf/2104.06892"]}
{"year":"2021","title":"Knowledge-enhanced neural grammar Induction","authors":["B Li - 2021"],"snippet":"… become the dominant approach. Thanks to the Transformer's superior capability of parallelism, PLMs can be efficiently trained on enormous raw text (eg, common crawl of the internet) containing billions of tokens. Surprisingly, although PLMs contain no …","url":["https://era.ed.ac.uk/bitstream/handle/1842/37896/BowenLi_2021.pdf?sequence=1"]}
{"year":"2021","title":"Label Imputation for Homograph Disambiguation: Theoretical and Practical Approaches","authors":["JM Seale - 2021"],"snippet":"Page 1. City University of New York (CUNY) CUNY Academic Works Dissertations, Theses, and Capstone Projects CUNY Graduate Center 9-2021 Label Imputation for Homograph Disambiguation: Theoretical and Practical Approaches Jennifer M. Seale …","url":["https://academicworks.cuny.edu/cgi/viewcontent.cgi?article=5591&context=gc_etds"]}
{"year":"2021","title":"Label-Descriptive Patterns and their Application to Characterizing Classification Errors","authors":["M Hedderich, J Fischer, D Klakow, J Vreeken - arXiv preprint arXiv:2110.09599, 2021"],"snippet":"State-of-the-art deep learning methods achieve human-like performance on many tasks, but make errors nevertheless. Characterizing these errors in easily interpretable terms gives insight into whether a model is prone to making systematic …","url":["https://arxiv.org/pdf/2110.09599"]}
{"year":"2021","title":"LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairs","authors":["C Schuhmann, R Vencu, R Beaumont, R Kaczmarczyk… - arXiv preprint arXiv …, 2021"],"snippet":"… To create image-text pairs, we parse through WAT files from Common Crawl and parse out all HTML IMG tags containing an alt-text attribute. We download the raw images from the parsed URLs with asynchronous requests using Trio and Asks libraries. …","url":["https://arxiv.org/pdf/2111.02114"]}
{"year":"2021","title":"LAMBERT: Layout-Aware Language Modeling for Information Extraction","authors":["Ł Garncarek, R Powalski, T Stanisławek, B Topolski…"],"snippet":"… tuning, respectively. The training was performed on a collection of PDFs extracted from Common Crawl containing a variety of documents (we randomly selected up to 10 documents from any single domain). The documents …","url":["https://www.researchgate.net/profile/Lukasz-Garncarek/publication/339374440_LAMBERT_Layout-Aware_language_Modeling_for_information_extraction/links/605c2fdda6fdccbfea04b14a/LAMBERT-Layout-Aware-language-Modeling-for-information-extraction.pdf"]}
{"year":"2021","title":"Language Embeddings for Typology and Cross-lingual Transfer Learning","authors":["D Yu, T He, K Sagae - arXiv preprint arXiv:2106.02082, 2021"],"snippet":"… lang_emb represents language embeddings from our proposed denoising autoencoder trained with language specific MUSE embeddings, using CommonCrawl text. Eng. lang_emb represents language embeddings trained …","url":["https://arxiv.org/pdf/2106.02082"]}
{"year":"2021","title":"Language Models are not Models of Language","authors":["C Veres - arXiv preprint arXiv:2112.07055, 2021"],"snippet":"… Also mentioned is the Common Crawl web content where \"... difficulty of cleaning and filtering the Common Crawl data ...\" which has low … Common Crawl data also contains duplicate lines which often signals highly repetitive \"boilerplate text\", which …","url":["https://arxiv.org/pdf/2112.07055"]}
{"year":"2021","title":"Language Models have a Moral Dimension","authors":["P Schramowski, C Turan, N Andersen, C Rothkopf… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Language Models have a Moral Dimension Patrick Schramowski, Cigdem Turan, Nico Andersen, Constantin Rothkopf & Kristian Kersting Abstract Artificial writing is permeating our lives due to recent advances in large …","url":["https://arxiv.org/pdf/2103.11790"]}
{"year":"2021","title":"Language Representation Models: An Overview","authors":["T Schomacker, M Tropmann-Frick - Entropy, 2021"],"snippet":"… Since C4 is based on the common crawl project, it not only contains book-like or article-like texts but a much wider variety of English texts. Despite its colossal size [11], p. 7, ensured that, by their used techniques, C4 only consists of reasonably clean …","url":["https://www.mdpi.com/1099-4300/23/11/1422/pdf"]}
{"year":"2021","title":"Language representations for computational argumentation","authors":["A Lauscher - 2021"],"snippet":"… CA Computational Argumentation. – , , , CBOW Continuous Bag of Words. – , , , CC Citation Contexts. CC- CommonCrawl- . , CCC Citation Context Identi cation. – CL Computational Linguistics. , , , , CLWE Cross-lingual Word Embe","url":["https://madoc.bib.uni-mannheim.de/60201/1/dissertation_lauscher.pdf"]}
{"year":"2021","title":"Language Transfer for Identifying Diagnostic Paragraphs in Clinical","authors":["L Di Liello, O Uryupina, A Moschitti - 2021"],"snippet":"This paper aims at uncovering the structure of clinical documents, in particular, identifying paragraphs describing “diagnosis” or “procedures”. We present transformer-based architectures for approaching this task in a monolingual setting (English) …","url":["http://ceur-ws.org/Vol-3033/paper56.pdf"]}
{"year":"2021","title":"LaoPLM: Pre-trained Language Models for Lao","authors":["N Lin, Y Fu, Z Yang, S Jiang - arXiv preprint arXiv:2110.05896, 2021"],"snippet":"… On the one hand, we utilize all the Lao data from the OSCAR corpus9, a humongous multilingual corpus whose texts all come from the Common Crawl corpus10. Suárez et al. [23] proposes the goclassy architecture to perform language …","url":["https://arxiv.org/pdf/2110.05896"]}
{"year":"2021","title":"Large Margin Training Improves Language Models for ASR","authors":["J Wang, J Huang, KW Church - ICASSP 2021-2021 IEEE International Conference on …, 2021"],"snippet":"… on enWiki+bookcorpus. We thus adopt the published model and skip the pretraining phase. On the other hand, GPT, as a realization of transformer decoder model, is pretrained on a larger corpus, common crawl. For a fair …","url":["https://ieeexplore.ieee.org/abstract/document/9414724/"]}
{"year":"2021","title":"Large-Scale Contextualised Language Modelling for Norwegian","authors":["A Kutuzov, J Barnes, E Velldal, L Øvrelid, S Oepen - arXiv preprint arXiv:2104.06546, 2021"],"snippet":"… models for Bokmål and Nynorsk). However, these models were trained on very modestly sized corpora of 20 million words for each language (randomly sampled from Wikipedia dumps and Common Crawl data). In a parallel effort …","url":["https://arxiv.org/pdf/2104.06546"]}
{"year":"2021","title":"Large-Scale Self-and Semi-Supervised Learning for Speech Translation","authors":["C Wang, A Wu, J Pino, A Baevski, M Auli, A Conneau - arXiv preprint arXiv …, 2021"],"snippet":"… We also make use of monolingual text data in the target language to further improve translation quality. Specifically, we train a language model (LM) on part of the CommonCrawl data3 that is in similar domains as CoVoST …","url":["https://arxiv.org/pdf/2104.06678"]}
{"year":"2021","title":"LARO: Language Agnostic Sentence Representations from finetuned RoBERTa⋆","authors":["AS Salvado - Language"],"snippet":"… The training resulted in the model approximating the meaning and syntax of all languages in the training data. The dataset which XLM-Roberta was trained on consists of 2.5TB CommonCrawl Data. The architecture of …","url":["https://users.informatik.haw-hamburg.de/~ubicomp/projekte/master2021-proj/soblechero.pdf"]}
{"year":"2021","title":"LaTr: Layout-Aware Transformer for Scene-Text VQA","authors":["AF Biten, R Litman, Y Xie, S Appalaraju, R Manmatha - arXiv preprint arXiv …, 2021"],"snippet":"We propose a novel multimodal architecture for Scene Text Visual Question Answering (STVQA), named Layout-Aware Transformer (LaTr). The task of STVQA requires models to reason over different modalities. Thus, we first investigate the …","url":["https://arxiv.org/pdf/2112.12494"]}
{"year":"2021","title":"LATS: Low-Resource Abstractive Text Summarization","authors":["CB van Yperen, F Frasincar, K Gruber - 2021"],"snippet":"… texts from 350 million Web pages. The dataset is based on the Common Crawl Web archive which is publicly available and contains scraped HTML files from which the markup has been removed. The HTML files have been …","url":["https://thesis.eur.nl/pub/56900/MSc_Thesis_27042021_CvanYperen.pdf"]}
{"year":"2021","title":"Learned Construction Grammars Converge Across Registers Given Increased Exposure","authors":["J Dunn, HT Madabushi - arXiv preprint arXiv:2110.05663, 2021"],"snippet":"This paper measures the impact of increased exposure on whether learned construction grammars converge onto shared representations when trained on data from different registers. Register influences the frequency of constructions, with …","url":["https://arxiv.org/pdf/2110.05663"]}
{"year":"2021","title":"Learning adaptive representations for entity recognition in the biomedical domain","authors":["I Lauriola, F Aiolli, A Lavelli, F Rinaldi - Journal of Biomedical Semantics, 2021"],"snippet":"Named Entity Recognition is a common task in Natural Language Processing applications, whose purpose is to recognize named entities in textual documents. Several systems exist to solve this task in the biomedical domain …","url":["https://jbiomedsem.biomedcentral.com/articles/10.1186/s13326-021-00238-0"]}
{"year":"2021","title":"Learning Mathematical Properties of Integers","authors":["M Ryskina, K Knight - arXiv preprint arXiv:2109.07230, 2021"],"snippet":"… Ours OEIS 14M 133,004 GloVe–840B–300D Common Crawl 840B 64,729 … (2019), we choose three embedding models pre-trained on English text corpora: • GloVe–840B–300D: 300-dimensional GloVe vectors trained …","url":["https://arxiv.org/pdf/2109.07230"]}
{"year":"2021","title":"Learning policy scheduling for text augmentation","authors":["S Li, X Ao, F Pan, Q He - Neural Networks, 2021"],"snippet":"Abstract When training deep learning models, data augmentation is an important technique to improve the performance and alleviate overfitting. In natural language processing (NLP), existing augmentation methods often use fixed strategies …","url":["https://www.sciencedirect.com/science/article/pii/S0893608021003907"]}
{"year":"2021","title":"Learning Sense-Specific Static Embeddings using Contextualised Word Embeddings as a Proxy","authors":["D Bollegala, Y Zhou - Proceedings of the 35th Pacific Asia Conference on …, 2021","Y Zhou, D Bollegala - arXiv preprint arXiv:2110.02204, 2021"],"snippet":"… context t . 3 Experiments 3.1 Experimental Setup In our experiments, we use the pretained GloVe3 em- beddings (Common Crawl with 840B tokens and 2.2M vocabulary) as the static word embeddings g(u) with p = 300. We use …","url":["https://aclanthology.org/2021.paclic-1.2.pdf","https://arxiv.org/pdf/2110.02204"]}
{"year":"2021","title":"Learning Sentiment Sentence Representation with Multiview Attention Model","authors":["Y Zhang, J Wang, X Zhang - Information Sciences, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0020025521005041"]}
{"year":"2021","title":"Learning Structured Neural Semantic Parsers","authors":["P Yin - 2021"],"snippet":"Page 1. Learning Structured Neural Semantic Parsers Pengcheng Yin CMU LTI 21 014 August 16, 2021 Language Technologies Institute School of Computer Science Carnegie Mellon University 5000 Forbes Ave., Pittsburgh, PA 15123 www.lti.cs.cmu.edu …","url":["https://www.lti.cs.cmu.edu/sites/default/files/yin%2C%20pengcheng%20-%20Thesis.pdf"]}
{"year":"2021","title":"Learning to Lemmatize in the Word Representation Space","authors":["J Lagus, A Klami - NoDaLiDa 2021, 2021"],"snippet":"… Data We validate the approach on Finnish language, using pretrained embeddings provided by the fastText library (Bojanowski et al., 2016). The embeddings were trained on Common Crawl and Wikipedia corpora and have dimensionality of d= 300 …","url":["https://ep.liu.se/ecp/178/ecp2021178.pdf#page=265"]}
{"year":"2021","title":"Learning to rank for Consumer Health Search","authors":["H Yang, X Liu, B Zheng, G Yang - CLEF 2021 Evaluation Labs and Workshop: Online …, 2021"],"snippet":"… Totally, the collection consists of over 5 million medical webpages from selected domains acquired from the CommonCrawl and other resources [4]. 3.3. Topics Totally 55 topics are used in the CLEF 2021 CHS task, and they are based on realistic search scenarios …","url":["http://ceur-ws.org/Vol-2936/paper-71.pdf"]}
{"year":"2021","title":"Learning To Retrieve Prompts for In-Context Learning","authors":["O Rubin, J Herzig, J Berant - arXiv preprint arXiv:2112.08633, 2021"],"snippet":"In-context learning is a recent paradigm in natural language understanding, where a large pre-trained language model (LM) observes a test instance and a few training examples as its input, and directly decodes the output without any update to its …","url":["https://arxiv.org/pdf/2112.08633"]}
{"year":"2021","title":"Leveraging Longitudinal Data for Personalized Prediction and Word Representations","authors":["C Welch - 2021"],"snippet":"Page 1. Leveraging Longitudinal Data for Personalized Prediction and Word Representations by Charles Welch A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Computer …","url":["https://deepblue.lib.umich.edu/bitstream/handle/2027.42/167971/cfwelch_1.pdf?sequence=1"]}
{"year":"2021","title":"Leveraging Transformers for Hate Speech Detection in Conversational Code-Mixed Tweets","authors":["ZM Farooqi, S Ghosh, RR Shah - arXiv preprint arXiv:2112.09986, 2021"],"snippet":"… • XLM-Roberta (XLM-R) [42] is a transformer-based masked language model pretrained on Common Crawl data having about 100 languages. It was proposed by Facebook and happened to be one of the best-performing transformer models for …","url":["https://arxiv.org/pdf/2112.09986"]}
{"year":"2021","title":"Lexicon Development for COVID-19-related Concepts Using Open-source Word Embedding Sources: An Intrinsic and Extrinsic Evaluation","authors":["S Parikh, A Davoudi, S Yu, C Giraldo, E Schriver… - JMIR Medical Informatics, 2021"],"snippet":"Background: Scientists are developing new computational methods and prediction models to better clinically understand COVID-19 prevalence, treatment efficacy, and patient outcomes. These efforts could be improved by leveraging …","url":["https://medinform.jmir.org/2021/2/e21679"]}
{"year":"2021","title":"Light gradient boosting machine-based phishing webpage detection model using phisher website features of mimic URLs","authors":["E Oram, PB Dash, B Naik, J Nayak, S Vimal, SK Nataraj - Pattern Recognition Letters, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0167865521003445"]}
{"year":"2021","title":"LILE: Look In-Depth before Looking Elsewhere--A Dual Attention Network using Transformers for Cross-Modal Information Retrieval in Histopathology Archives","authors":["D Maleki, H Tizhoosh - 2021"],"snippet":"The volume of available data has grown dramatically in recent years in many applications. Furthermore, the age of networks that used multiple modalities separately has practically ended. Therefore, enabling bidirectional cross-modality …","url":["https://openreview.net/pdf?id=ZmsElUaQ0Xm"]}
{"year":"2021","title":"Limitations of Knowledge Distillation for Zero-shot Transfer Learning","authors":["S Soltan, AA AI, H Khan, W Hamza"],"snippet":"… However, they focused on distillation during pretraining and evaluate their model only on XNLI from XTREME benchmarks. Since they used common crawl data for pretraining and a different architecture, their models are not directly comparable to ours. 3 Method …","url":["https://assets.amazon.science/3b/46/17df4724436d850c46fcb41b1569/limitations-of-knowledge-distillation-for-zero-shot-transfer-learning.pdf"]}
{"year":"2021","title":"LIMSI@ MLIA","authors":["SA Rauf, F Yvon - 2020"],"snippet":"… In addition, we used the Cochrane bilingual parallel corpus [4]3 and the Taus Corona Crisis corpus.4. For out-of-domain corpus, we used the parallel corpora provided by the WMT14 campaign, which included Gigafr-en …","url":["http://eval.covid19-mlia.eu/meetings/round1/report/20210114-limsi.pdf"]}
{"year":"2021","title":"Lingua Custodia's participation at the WMT 2021 Machine Translation using Terminologies shared task","authors":["M Ailem, J Liu, R Qader - arXiv preprint arXiv:2111.02120, 2021"],"snippet":"This paper describes Lingua Custodia's submission to the WMT21 shared task on machine translation using terminologies. We consider three directions, namely English to French, Russian, and Chinese. We rely on a Transformer-based …","url":["https://arxiv.org/pdf/2111.02120"]}
{"year":"2021","title":"Linking place records using multi-view encoders","authors":["V Cousseau, L Barbosa - Neural Computing and Applications, 2021"],"snippet":"… For address embeddings, we utilized case-sensitive embeddings trained with the FastText model [4] on top of a Wikipedia and Common Crawl corpus [22]. The reasoning behind the different embedding sources for names and …","url":["https://link.springer.com/article/10.1007/s00521-021-05932-9"]}
{"year":"2021","title":"LIORI at SemEval-2021 Task 2: Span Prediction and Binary Classification approaches to Word-in-Context Disambiguation","authors":["A Davletov, N Arefyev, D Gordeev, A Rey - Proceedings of the Fifteenth Workshop on …, 2021"],"snippet":"… tokens. XLM-R was trained on a cleaned two-terabyte CommonCrawl Corpus in 100 languages. A new promising approach to language task modelling is treating any natural language task as a question answering problem …","url":["https://aclanthology.org/2021.semeval-1.103.pdf"]}
{"year":"2021","title":"LISN@ WMT 2021","authors":["J Xu, SA Rauf, MQ Pham, F Yvon - 6th Conference on Statistical Machine Translation, 2021"],"snippet":"This paper describes LISN's submissions to two shared tasks at WMT'21. For the biomedical translation task, we have developed resource-heavy systems for the English-French language pair, using both out-ofdomain and in-domain corpora. The …","url":["https://hal.archives-ouvertes.fr/hal-03430610/document"]}
{"year":"2021","title":"LMMS Reloaded: Transformer-based Sense Embeddings for Disambiguation and Beyond","authors":["D Loureiro, AM Jorge, J Camacho-Collados - arXiv preprint arXiv:2105.12449, 2021"],"snippet":"Page 1. LMMS Reloaded: Transformer-based Sense Embeddings for Disambiguation and Beyond Daniel Loureiroa,∗, Alıpio Mário Jorgea, Jose Camacho-Colladosb aLIAAD - INESC TEC, Dept. of Computer Science, FCUP …","url":["https://arxiv.org/pdf/2105.12449"]}
{"year":"2021","title":"LNN-EL: A Neuro-Symbolic Approach to Short-text Entity Linking","authors":["H Jiang, S Gurajada, Q Lu, S Neelam, L Popa, P Sen… - arXiv preprint arXiv …, 2021"],"snippet":"… We also employ a suite of pretrained or custom trained neural language models to compute the similarity of mi and eij. Pre-trained Embedding Models. These include SpaCy's semantic similarity6 function that uses Glove …","url":["https://arxiv.org/pdf/2106.09795"]}
{"year":"2021","title":"Log Parsing and Template Extraction Using Neueral Sequence-To-Sequence Models","authors":["K Pawlikowski - 2021"],"snippet":"Abstract System log messages are produced by software systems as their code executes. The contents of these log messages vary across systems and are defined by the software developers of the system. Log messages can be captured and …","url":["https://search.proquest.com/openview/6e178e5359e69ed4aa85193bc726d8ac/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Looking for Semantic Similarity: What a Vector Space Model of Semantics Can Tell Us About Attention in Real-world Scenes","authors":["TR Hayes, JM Henderson - 2021"],"snippet":"… ing the semantic vectors from Word2vec (Mikolov, Chen, Corrado, & Dean, 2013) and GloVe 1.2 (Pennington, Socher, & Manning, 2014) that learn how words are associated with each other from large text corpora (ie, Google …","url":["https://osf.io/wsyz9/download?format=pdf"]}
{"year":"2021","title":"Low Frequency Names Exhibit Bias and Overfitting in Contextualizing Language Models","authors":["R Wolfe, A Caliskan - arXiv preprint arXiv:2110.00672, 2021"],"snippet":"Page 1. Low Frequency Names Exhibit Bias and Overfitting in Contextualizing Language Models Robert Wolfe University of Washington rwolfe3@uw.edu Aylin Caliskan University of Washington aylin@uw.edu Abstract We use …","url":["https://arxiv.org/pdf/2110.00672"]}
{"year":"2021","title":"LSI_UNED at CLEF eHealth2021: Exploring the effects of transfer learning in negation detection and entity recognition in clinical texts","authors":["H Fabregat, A Duque, L Araujo, J Martinez-Romo - CLEF eHealth, 2021"],"snippet":"… Word embeddings: Two different pre-trained word embeddings from different sources are used for text representation. On the one hand, we use general domain Spanish 100dimensional word embeddings in …","url":["http://ceur-ws.org/Vol-2936/paper-65.pdf"]}
{"year":"2021","title":"M6: A Chinese Multimodal Pretrainer","authors":["J Lin, R Men, A Yang, C Zhou, M Ding, Y Zhang… - arXiv preprint arXiv …, 2021"],"snippet":"… Xu et al. [45] released a 100GB corpus named CLUECorpus2020, which is re- trieved from the multilingual Common Crawl dataset … Forum discussion Plain-text - 8.7 39.0 223.1 - 18.0 Common Crawl Plain-text - 40.3 108.7 370.7 - 83.3 …","url":["https://arxiv.org/pdf/2103.00823"]}
{"year":"2021","title":"M6: Multi-Modality-to-Multi-Modality Multitask Mega-transformer for Unified Pretraining","authors":["J Lin, R Men, A Yang, C Zhou, Y Zhang, P Wang… - 2021"],"snippet":"… Xu et al. [45] released a 100𝐺𝐵 corpus named CLUECorpus2020, which is retrieved from the multilingual Common Crawl dataset … Forum discussion Plain-text - 8.7 39.0 223.1 - 18.0 Common Crawl Plain-text - 40.3 108.7 370.7 - 83.3 …","url":["https://keg.cs.tsinghua.edu.cn/jietang/publications/KDD21-Lin-et-al-M6.pdf"]}
{"year":"2021","title":"Machine Habitus: Toward a Sociology of Algorithms","authors":["M Airoldi - 2021"]}
{"year":"2021","title":"Machine Learning and Economic Models to Enable Risk-Informed Condition Based Maintenance of a Nuclear Plant Asset","authors":["V Agarwal, A Manjunatha, JA Smith, AV Gribok… - 2021"],"snippet":"Page 1. INL/EXT-21-61984 Revision 0 Industry Funding Opportunity Announcement Award Machine Learning and Economic Models to Enable Risk-Informed Condition Based Maintenance of a Nuclear Plant Asset March …","url":["https://www.osti.gov/servlets/purl/1770866"]}
{"year":"2021","title":"Machine Learning and Security Studies","authors":["S Dorsey - 2020"],"snippet":"Page 1. Machine Learning and Security Studies by Spencer Dorsey Department of Political Science Duke University Date: Approved: Scott de Marchi, Advisor Peter Feaver, Advisor Laia Balcells Erik Wibbels Dissertation submitted …","url":["https://dukespace.lib.duke.edu/dspace/bitstream/handle/10161/22158/Dorsey_duke_0066D_15915.pdf?sequence=1"]}
{"year":"2021","title":"Machine learning for financial transaction classification across companies using character-level word embeddings of text fields","authors":["RK Jørgensen, C Igel - Intelligent Systems in Accounting, Finance and …"],"snippet":"Abstract An important initial step in accounting is mapping financial transfers to the corresponding accounts. We devised machine-learning-based systems that automate this process. They use word em...","url":["https://onlinelibrary.wiley.com/doi/full/10.1002/isaf.1500"]}
{"year":"2021","title":"Machine Learning Meets Natural Language Processing--The story so far","authors":["NI Galanis, P Vafiadis, KG Mirzaev, GA Papakostas - arXiv preprint arXiv:2104.10213, 2021"],"snippet":"… in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4938–4947, 2020. 42. “Common Crawl”, Accessed: Mar. 24, 2021.[Online]. Available: https://commoncrawl.org/ 43. “Kaggle”, Accessed: Mar. 24, 2021.[Online] …","url":["https://arxiv.org/pdf/2104.10213"]}
{"year":"2021","title":"Machine Learning with Applications","authors":["N Kanakaris, N Giarelis, I Siachos, N Karacapilidis","NK Corrêa, S Falk, S Fatimah, A Sen, N De Oliveira","S Casola, I Lauriola, A Lavelli"],"snippet":"This paper employs techniques and algorithms from the fields of natural language processing, graph representation learning and word embeddings to assist project managers in the task of personnel selection. To do so, our approach initially …","url":["https://www.researchgate.net/profile/Nicholas-Correa-2/publication/380496051_TeenyTinyLlama_Open-source_tiny_language_models_trained_in_Brazilian_Portuguese/links/66474976479366623af3b128/TeenyTinyLlama-Open-source-tiny-language-models-trained-in-Brazilian-Portuguese.pdf","https://www.researchgate.net/profile/Nikos-Kanakaris/publication/356403657_Making_personnel_selection_smarter_through_word_embeddings_A_graph-based_approach/links/61a4ac35d7d1af224b2be56e/Making-personnel-selection-smarter-through-word-embeddings-A-graph-based-approach.pdf","https://www.researchgate.net/profile/Silvia-Casola/publication/361083726_Pre-trained_transformers_an_empirical_comparison/links/62a3370f55273755ebe1d59b/Pre-trained-transformers-an-empirical-comparison.pdf"]}
{"year":"2021","title":"Machine Translation Customization via Automatic Training Data Selection from the Web","authors":["T Vu, A Moschitti"],"snippet":"… 5 Fig. 1: Customization process of MT Systems Corpus Sent. (MM) News Commentary v13 0.3 Rapid (press releases) 1.3 Common Crawl 1.9 Europarl v7 2.4 ParaCrawl (Zipporah) 40.6 ParaCrawl (BiCleaner) 27.7 …","url":["https://assets.amazon.science/ed/9f/3bc994bc4be7bc53449bda397d83/machine-translation-customization-via-automatic-training-data-selection-from-the-web.pdf"]}
{"year":"2021","title":"Majority Voting with Bidirectional Pre-translation Improves Bitext Retrieval","authors":["A Jones, DT Wijaya - arXiv preprint arXiv:2103.06369, 2021"],"snippet":"Page 1. Majority Voting with Bidirectional Pre-translation Improves Bitext Retrieval Alex Jones Dartmouth College alexander.g.jones.23@dartmouth.edu Derry Tanti Wijaya Boston University wijaya@bu.edu Abstract Obtaining …","url":["https://arxiv.org/pdf/2103.06369"]}
{"year":"2021","title":"Making personnel selection smarter through word embeddings: A graph-based approach","authors":["N Kanakaris, N Giarelis, I Siachos, N Karacapilidis - Machine Learning with …, 2021"],"snippet":"… GoogleNews using Word2Vec and Common Crawl using GloVe are some of the most favored public pre-trained word embeddings. For an in-depth analysis as far as the differences and use-cases of various pre-trained word embeddings are …","url":["https://www.sciencedirect.com/science/article/pii/S2666827021001079"]}
{"year":"2021","title":"Maps Search Misspelling Detection Leveraging Domain-Augmented Contextual Representations","authors":["Y Li - arXiv preprint arXiv:2108.06842, 2021"],"snippet":"… The second two models are trained with 42B Glove (Common Crawl) vectors, with embedding dimension 300, hidden state di- mension 50 and the same vocabulary. Our motivation of using a Page 4. KDD WIT '21, Aug …","url":["https://arxiv.org/pdf/2108.06842"]}
{"year":"2021","title":"Marked Attribute Bias in Natural Language Inference","authors":["H Dawkins"],"snippet":"… Otherwise, we apply the method to the 2Taken as the GloVe embeddings trained on the Common Crawl corpus for 840B tokens; available at https://nlp.stanford.edu/projects/glove/. Results were not found to vary …","url":["https://aclanthology.org/2021.findings-acl.369.pdf"]}
{"year":"2021","title":"MarkupLM: Pre-training of Text and Markup Language for Visually-rich Document Understanding","authors":["J Li, Y Xu, L Cui, F Wei - arXiv preprint arXiv:2110.08518, 2021"],"snippet":"… Equipped with the existing webpage datasets Common Crawl (CC)3, we pre-train MarkupLM with large-scale unlabeled HTML data and evaluate the pre-trained models on web-based structural reading comprehension and information extraction tasks. …","url":["https://arxiv.org/pdf/2110.08518"]}
{"year":"2021","title":"MasakhaNER: Named Entity Recognition for African Languages","authors":["DI Adelani, J Abbott, G Neubig, D D'souza, J Kreutzer… - arXiv preprint arXiv …, 2021"],"snippet":"… XLMR XLMR (Conneau et al., 2020) was trained on 100 languages including Amharic, Hausa, and Swahili. The major differences be tween XLMR and mBERT are (1) XLMR was trained on Common Crawl while mBERT was …","url":["https://arxiv.org/pdf/2103.11811"]}
{"year":"2021","title":"Masked Language Model Entity Matching for Cultural Heritage Data","authors":["D Piché, A Zouaq, M Gagnon, L Font - 2021"],"snippet":"… CamemBERT [10] is trained on the French portion of the OSCAR corpus, a pre-filtered version of Common Crawl, while FlauBERT [8] trains on a combination of French sets, including WMT19, OPUS, Wikimedia and Project Gutenberg. 3 Data overview …","url":["http://ceur-ws.org/Vol-2949/paper5.pdf"]}
{"year":"2021","title":"Mastering Transformers: Build state-of-the-art models from scratch with advanced natural language processing techniques","authors":["S Yildirim, M Asgari-Chenaghlu - 2021"],"snippet":"Page 1. Mastering Transformers Build state-of-the-art models from scratch with advanced natural language processing techniques 0001) il -0 1 0 -0 -O 0 - -0 1 -0 - -0 1 1. -0 -0 -0 -0 -0 -0 1 1 0 010 1 0 1 0 1 01100110 0001101010 …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=A9Y6EAAAQBAJ&oi=fnd&pg=PP1&dq=commoncrawl&ots=vd7ymgpejf&sig=Vl9kWAVZADIBu0pKOCgd-uJ7Xw8"]}
{"year":"2021","title":"Matches Made in Heaven: Toolkit and Large-Scale Datasets for Supervised Query Reformulation","authors":["N Arabzadeh, A Bigdeli, S Seyedsalehi, M Zihayat…"],"snippet":"Page 1. Matches Made in Heaven: Toolkit and Large-Scale Datasets for Supervised Query Reformulation Negar Arabzadeh University of Waterloo narabzad@uwaterloo. ca Amin Bigdeli Ryerson University abigdeli@ryerson.ca …","url":["http://ls3.rnet.ryerson.ca/wiki/images/a/ab/Matches-made-in-heaven.pdf"]}
{"year":"2021","title":"MDAPT: Multilingual Domain Adaptive Pretraining in a Single Model","authors":["RK Jørgensen, M Hartmann, X Dai, D Elliott - arXiv preprint arXiv:2109.06605, 2021"],"snippet":"Page 1. MDAPT: Multilingual Domain Adaptive Pretraining in a Single Model Rasmus Kær Jørgensen1,2 and Mareike Hartmann3∗ and Xiang Dai1 and Desmond Elliott1 1University of Copenhagen, Denmark 2PricewaterhouseCoopers …","url":["https://arxiv.org/pdf/2109.06605"]}
{"year":"2021","title":"Measuring associational thinking through word embeddings","authors":["C Periñán-Pascual - Artificial Intelligence Review, 2021"],"snippet":"The development of a model to quantify semantic similarity and relatedness between words has been the major focus of many studies in various fields, eg p.","url":["https://link.springer.com/article/10.1007/s10462-021-10056-6"]}
{"year":"2021","title":"Measuring Gender Bias in Word Embeddings for Russian Language","authors":["AS Pestova"],"snippet":"… The effect size is calculated as Cohen's d [6]: mean𝑥𝑥∈𝑋𝑋 𝑐𝑐(𝑥𝑥𝑎 𝐴𝐴𝑎 𝐵𝐵) - mean𝑦𝑦∈𝑌𝑌 𝑐𝑐(𝑦𝑦𝑎 𝐴𝐴𝑎 𝐵𝐵) std - dev𝑤𝑤∈𝑋𝑋∪𝑌𝑌 𝑐𝑐(𝑤𝑤𝑎 𝐴𝐴𝑎 𝐵𝐵) (4) 3https://tatianashavrina.github.io/taiga\\_site …","url":["http://www.dialog-21.ru/media/5322/pestovaas047.pdf"]}
{"year":"2021","title":"Memes in the Wild: Assessing the Generalizability of the Hateful Memes Challenge Dataset","authors":["HR Kirk, Y Jun, P Rauba, G Wachtel, R Li, X Bai… - arXiv preprint arXiv …, 2021"],"snippet":"… F T-SNE Text Embeddings The meme-level embeddings are calculated by (i) extracting a 300-dimensional embedding for each word in the meme, using fastText embeddings trained on Wikipedia and Common Crawl; (ii) …","url":["https://arxiv.org/pdf/2107.04313"]}
{"year":"2021","title":"Mengzi: Towards Lightweight yet Ingenious Pre-trained Models for Chinese","authors":["Z Zhang, H Zhang, K Chen, Y Guo, J Hua, Y Wang… - arXiv preprint arXiv …, 2021"],"snippet":"Although pre-trained models (PLMs) have achieved remarkable improvements in a wide range of NLP tasks, they are expensive in terms of time and resources. This calls for the study of training more efficient models with less computation but still …","url":["https://arxiv.org/pdf/2110.06696"]}
{"year":"2021","title":"MetaXL: Meta Representation Transformation for Low-resource Cross-lingual Learning","authors":["M Xia, G Zheng, S Mukherjee, M Shokouhi, G Neubig… - arXiv preprint arXiv …, 2021"],"snippet":"… The Hausdorff distance between the source and target representations drops from 0.57 to 0.20 with F1 score improvement from 74.07 to 78.15. Wikipedia and XLM-R (Conneau et al., 2020) is pre-trained on 100 languages with CommonCrawl Corpora …","url":["https://arxiv.org/pdf/2104.07908"]}
{"year":"2021","title":"Metric Learning in Multilingual Sentence Similarity Measurement for Document Alignment","authors":["C Rajitha, L Piyarathne, D Sachintha, S Ranathunga - arXiv preprint arXiv …, 2021"],"snippet":"… applied to the training data. XLM-R is trained with common crawl data from 100 languages, and has shown to outperform mBERT on multiple Natural Language Processing tasks. 3 Metric Learning Unsupervised metrics such as …","url":["https://arxiv.org/pdf/2108.09495"]}
{"year":"2021","title":"MFAQ: a Multilingual FAQ Dataset","authors":["M De Bruyn, E Lotfi, J Buhmann, W Daelemans - arXiv preprint arXiv:2109.12870, 2021"],"snippet":"… provides an open repository of the web.5 Common Crawl's complete web archive consists of petabytes of data collected over 10 years of web crawling (Ortiz Suárez et al., 2020). The repository is organized in monthly bucket …","url":["https://arxiv.org/pdf/2109.12870"]}
{"year":"2021","title":"Mind the Gap: Assessing Temporal Generalization in Neural Language Models","authors":["A Lazaridou, A Kuncoro, E Gribovskaya, D Agrawal… - Advances in Neural …, 2021"],"snippet":"Abstract Our world is open-ended, non-stationary, and constantly evolving; thus what we talk about and how we talk about it change over time. This inherent dynamic nature of language contrasts with the current static language modelling …","url":["https://proceedings.neurips.cc/paper/2021/file/f5bf0ba0a17ef18f9607774722f5698c-Paper.pdf"]}
{"year":"2021","title":"Minimum Bayes Risk Decoding with Neural Metrics of Translation Quality","authors":["M Freitag, D Grangier, Q Tan, B Liang - arXiv preprint arXiv:2111.09388, 2021"],"snippet":"This work applies Minimum Bayes Risk (MBR) decoding to optimize diverse automated metrics of translation quality. Automatic metrics in machine translation have made tremendous progress recently. In particular, neural metrics, fine-tuned on …","url":["https://arxiv.org/pdf/2111.09388"]}
{"year":"2021","title":"Mining emotion-aware sequential rules at user-level from micro-blogs","authors":["MP Skenduli, M Biba, C Loglisci, M Ceci, D Malerba - Journal of Intelligent Information …, 2021"],"snippet":"… the fine-grained emotion analysis task. Based on the obtained results, we decide to use multi-lingual word embeddings, pre-trained on Common Crawl and Wikipedia corpora using fastText. These models were trained using …","url":["https://link.springer.com/article/10.1007/s10844-021-00647-8"]}
{"year":"2021","title":"Misinformation Retrieval","authors":["S Rizvi - 2021"],"snippet":"… 34 4.1 Excerpt from a correct document from the Common Crawl Collection with fact used to validate tweets in blue. . . . . 43 viii Page 9 … for a given topic. The corpus used for both tasks was the common crawl corpus consisting of news articles online …","url":["https://uwspace.uwaterloo.ca/bitstream/handle/10012/17609/Rizvi_Saira.pdf?sequence=3&isAllowed=y"]}
{"year":"2021","title":"Mitigating the Position Bias of Transformer Models in Passage Re-Ranking","authors":["S Hofstätter, A Lipani, S Althammer, M Zlabinger… - arXiv preprint arXiv:2101.06980"],"snippet":"… For the Transformer layers in TK we evaluate 2 layers each with 16 attention heads with size 32 and a feed-forward dimension of 100. For kernel-pooling we set the number 3 github.com/microsoft/BlingFire 4 42B CommonCrawl …","url":["https://arxiv.org/pdf/2101.06980"]}
{"year":"2021","title":"MKPM: Multi keyword-pair matching for natural language sentences","authors":["X Lu, Y Deng, T Sun, Y Gao, J Feng, X Sun, R Sutcliffe - Applied Intelligence, 2021"],"snippet":"Sentence matching is widely used in various natural language tasks, such as natural language inference, paraphrase identification and question answering. F.","url":["https://link.springer.com/article/10.1007/s10489-021-02306-5"]}
{"year":"2021","title":"MLLP-VRAIN Spanish ASR Systems for the Albayzin-RTVE 2020 Speech-To-Text Challenge","authors":["J Jorge, A Giménez, P Baquero-Arnal…"],"snippet":"… El Periódico [28] 2677 46637 291 Common Crawl [29] 1719 41792 486 Internal: parliamentary data 1361 35170 126 … [27] “Eldiario.es,” https://www.eldiario.es/. [28] “ElPeriodico.com,” https://www.elperiodico …","url":["https://www.isca-speech.org/archive/IberSPEECH_2021/pdfs/25.pdf"]}
{"year":"2021","title":"mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models","authors":["R Ri, I Yamada, Y Tsuruoka - arXiv preprint arXiv:2110.08151, 2021"],"snippet":"… 2019) trained with CommonCrawl data from 100 languages, and the parameters in common (eg, the weights of the transformer encoder and the word embeddings) are initialized using the checkpoint from the Transformers library.The size of the …","url":["https://arxiv.org/pdf/2110.08151"]}
{"year":"2021","title":"Mobile Testing Framework Exploiting Machine Learning and NLP","authors":["L ARDITO, M MORISIO, R COPPOLA - 2021"],"snippet":"… the NLP step. Bert-Base, Multilingual Cased for BERT feature vector, GloVe Common Crawl 300 dimensional model for Glove, and lastly Fasttext 300 dimensional pre-trained model for Fasttext vector was used. On the android …","url":["https://webthesis.biblio.polito.it/19106/1/tesi.pdf"]}
{"year":"2021","title":"Model Averaging in Distributed Machine Learning: A Case Study with Apache Spark","authors":["Y Guo, Z Zhang, J Jiang, W Wu, C Zhang, B Cui, J Li"],"snippet":"Page 1. Noname manuscript No. (will be inserted by the editor) Model Averaging in Distributed Machine Learning: A Case Study with Apache Spark Yunyan Guo · Zhipeng Zhang · Jiawei Jiang · Wentao Wu · Ce Zhang · Bin …","url":["http://pages.cs.wisc.edu/~wentaowu/papers/vldbj-model-avg.pdf"]}
{"year":"2021","title":"Modeling Endorsement for Multi-Document Abstractive Summarization","authors":["L Lebanoff, B Wang, Z Feng, F Liu - arXiv preprint arXiv:2110.07844, 2021"],"snippet":"A crucial difference between singleand multi-document summarization is how salient content manifests itself in the document(s). While such content may appear at the beginning of a single document, essential information is frequently reiterated in a …","url":["https://arxiv.org/pdf/2110.07844"]}
{"year":"2021","title":"Modeling Global and Local Interactions for Online Conversation Recommendation","authors":["X Zeng, J Li, L Wang, KF Wong - ACM Transactions on Information Systems (TOIS), 2021"],"snippet":"… For the parameters with respect to local interactions, we initialize the word embedding layer with 200-dimensional Glove embedding [66], where the Twitter version is used for our Twitter dataset, and the Common Crawl version is applied on …","url":["https://dl.acm.org/doi/full/10.1145/3473970"]}
{"year":"2021","title":"Modern-Day Shakespeare: Training Set Experiments with a Generative Pre-Trained Transformer-Best Paper","authors":["R Sheverack - 2021"],"snippet":"Page 1. 1 Modern-Day Shakespeare: Training Set Experiments with a Generative Pre-Trained Transformer Applied Project Final Report By Roksolana Maria Sheverack Spring 2021 A paper submitted in partial …","url":["http://archive.nyu.edu/bitstream/2451/62802/2/R.Sheverack%20-%20Final%20Project%20Report%20-%20Modern-Day%20Shakespeare%3ATraining%20Set%20Experiments%20with%20a%20Generative%20Pre-Trained%20Transformer.pdf"]}
{"year":"2021","title":"Monitoring COVID-19 pandemic through the lens of social media using natural language processing and machine learning","authors":["Y Liu, C Whitfield, T Zhang, A Hauser, T Reynolds… - Health Information Science …, 2021"],"snippet":"… The custom NER model that we trained was based on spaCy's multi-task, OntoNotes-trained Convolutional Neural Network which uses GloVe vectors that were trained using Common Crawl [37] corpus. Topic modeling. Topic …","url":["https://link.springer.com/article/10.1007/s13755-021-00158-4"]}
{"year":"2021","title":"Monolingual and Cross-Lingual Acceptability Judgments with the Italian CoLA corpus","authors":["D Trotta, R Guarasci, E Leonardelli, S Tonelli - arXiv preprint arXiv:2109.12053, 2021"],"snippet":"… Each sentence is represented as a sequence of word embeddings, obtained with the Italian model of FastText (Grave et al., 2018) trained on Common Crawl and Wikipedia with size 300.4 The network is implemented with Keras …","url":["https://arxiv.org/pdf/2109.12053"]}
{"year":"2021","title":"Moral rights","authors":["NB Eve"],"snippet":"… offer other services, including: CC Search, which makes openly licensed material easier to discover and use. As it makes use of the Common Crawl dataset, creators do not have to apply to be included (“About CC Search”) …","url":["http://sharedspaces.usask.ca/ipandcopyright"]}
{"year":"2021","title":"More Data and New Tools. Advances in Parsing the Index Thomisticus Treebank","authors":["F Gamba, M Passarotti, P Ruffolo - Proceedings http://ceur-ws. org ISSN, 2021"],"snippet":"Abstract This paper investigates the recent advances in parsing the Index Thomisticus Treebank, which encompasses Medieval Latin texts by Thomas Aquinas. The research focuses on two types of variables. On the one hand, it examines the …","url":["http://ceur-ws.org/Vol-2989/long_paper20.pdf"]}
{"year":"2021","title":"More Than Reading Comprehension: A Survey on Datasets and Metrics of Textual Question Answering","authors":["Y Bai, DZ Wang - arXiv preprint arXiv:2109.12264, 2021"],"snippet":"… Long-form QA ELI5[16] 2019 EN Web. Common Crawl Reddit Reddit Free-form ROUG-1,2,L MASH-QA[17] 2020 EN Health WebMD QA section QA section Span of words F1, EM NLQuAD[18] 2021 EN News BBC …","url":["https://arxiv.org/pdf/2109.12264"]}
{"year":"2021","title":"Morph Call: Probing Morphosyntactic Content of Multilingual Transformers","authors":["V Mikhailov, O Serikov, E Artemova - arXiv preprint arXiv:2104.12847, 2021"],"snippet":"… languages. D-BERT (Sanh et al., 2019) or DistilBERT is a 6-layer distilled version of M-BERT model. XLM-R (Conneau et al., 2019) was pre-trained over filtered CommonCrawl data in 100 languages (Wenzek et al., 2019). MiniLM …","url":["https://arxiv.org/pdf/2104.12847"]}
{"year":"2021","title":"Morphological Skip-Gram: Replacing FastText characters n-gram with morphological knowledge","authors":["TD Bispo, FAO Santos, HT Macedo, C Zanchettin - Inteligencia Artificial, 2021"],"snippet":"Page 1. Inteligencia Artificial, 24(67), 1-17 doi: 10.4114/intartif.vol24iss67pp1-17 INTELIGENCIA ARTIFICIAL http://journal.iberamia.org/ Morphological Skip-Gram: Replacing FastText characters n-gram with morphological knowledge …","url":["https://www.journal.iberamia.org/index.php/intartif/article/download/588/131"]}
{"year":"2021","title":"Multi-granular Legal Topic Classification on Greek Legislation","authors":["C Papaloukas, I Chalkidis, K Athinaios, DA Pantazi… - arXiv preprint arXiv …, 2021"],"snippet":"… learning rate. Based on this study, Conneau et al. (2020) proposed the XLMRoBERTa model, which supports 100 different languages (Greek included) and is trained on 2.5TB of filtered Common Crawl data. GREEK-BERT …","url":["https://arxiv.org/pdf/2109.15298"]}
{"year":"2021","title":"Multi-Granularity Contrasting for Cross-Lingual Pre-Training","authors":["S Li, P Yang, F Luo, J Xie"],"snippet":"… Following Conneau et al. (2019), we re- construct Common-Crawl Corpus to obtain monolingual training datasets while the bilingual data is obtained from the OPUS website 5. We also conduct up/down-sample for all pre-training data with a smoothing parameter …","url":["https://aclanthology.org/2021.findings-acl.149.pdf"]}
{"year":"2021","title":"Multi-level Embeddings for Processing Arabic Social Media Contents","authors":["L Moudjari, F Benamara, K Akli-Astouati - Computer Speech & Language, 2021"],"url":["https://www.sciencedirect.com/science/article/pii/S0885230821000474"]}
{"year":"2021","title":"Multi-Level Sentiment Analysis of Product Reviews Based on Grammar Rules","authors":["HD Nguyen, T Le, K Tran, ST Luu, SN Hoang, HT Phan - New Trends in Intelligent …, 2021"],"snippet":"… BERT is a transfer learning method which is trained on large language corpus such as Wikipedia and Common Crawl in different languages, then applied to specific tasks by fine tuning the model [7]. The appearance of BERT makes …","url":["https://ebooks.iospress.nl/volumearticle/57466"]}
{"year":"2021","title":"Multi-task Learning Using a Combination of Contextualised and Static Word Embeddings for Arabic Sarcasm Detection and Sentiment Analysis","authors":["AI Alharbi, M Lee - Proceedings of the Sixth Arabic Natural Language …, 2021"],"snippet":"… al., 2017). The Ara2Vec word embeddings is a well-known open-source tool made up of six Arabic models. Data was obtained from Wikipedia, Twitter and Common Crawl (webpage crawl data) to train Ara2Vec. Al- though it …","url":["https://www.aclweb.org/anthology/2021.wanlp-1.39.pdf"]}
{"year":"2021","title":"MultiEmo: Multilingual, Multilevel, Multidomain Sentiment Analysis Corpus of Consumer Reviews","authors":["K Kanclerz"],"snippet":"… The first model, Unsupervised Cross-lingual Representation Learning at Scale (XLM-RoBERTa), is a large multillingual language model, trained on 2.5TB of filtered CommonCrawl data, using self-supervised training techniques to achieve state-of-the-art performance …","url":["https://www.iccs-meeting.org/archive/iccs2021/papers/127430291.pdf"]}
{"year":"2021","title":"MultiEURLEX--A multi-lingual and multi-label legal document classification dataset for zero-shot cross-lingual transfer","authors":["I Chalkidis, M Fergadiotis, I Androutsopoulos - arXiv preprint arXiv:2109.00904, 2021"],"snippet":"… It is pretrained on Common Crawl with a vocabulary of 250k sub-words shared across languages … (2021) released a multilingual variant of T5 (Raffel et al., 2020), an encoder-decoder TRANSFORMER-based model …","url":["https://arxiv.org/pdf/2109.00904"]}
{"year":"2021","title":"Multilingual and Zero-Shot is Closing in on Monolingual Web Register Classification","authors":["S Rönnqvist, V Skantsi, M Oinonen, V Laippala - NoDaLiDa 2021, 2021"],"snippet":"… The English CORE is based on unrestricted search queries of extremely frequent n-grams, while the other datasets are randomly sampled from the 2017 CoNLL Shared Task datasets, originally drawn from Common Crawl (Ginter et al., 2017) …","url":["https://ep.liu.se/ecp/178/ecp2021178.pdf#page=173"]}
{"year":"2021","title":"Multilingual Autoregressive Entity Linking","authors":["N De Cao, L Wu, K Popat, M Artetxe, N Goyal… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Multilingual Autoregressive Entity Linking Nicola De Cao1,2, Ledell Wu1, Kashyap Popat1, Mikel Artetxe1, Naman Goyal1, Mikhail Plekhanov1, Luke Zettlemoyer1,3, Nicola Cancedda1, Sebastian Riedel1,4, Fabio Petroni1 …","url":["https://arxiv.org/pdf/2103.12528"]}
{"year":"2021","title":"Multilingual Counter Narrative Type Classification","authors":["YL Chung, M Guerini, R Agerri - arXiv preprint arXiv:2109.13664, 2021"],"snippet":"… The XLM-R language model (Conneau et al., 2020), pre-trained on CommonCrawl for 100 languages, reports strong performance on several cross-lingual downstream tasks such as natural language inference or …","url":["https://arxiv.org/pdf/2109.13664"]}
{"year":"2021","title":"Multilingual discussion summarisation in social networks","authors":["TN Andreevic - 2021"],"snippet":"Page 1. Санкт–Петербургский государственный университет Тарасов Никита Андреевич Выпускная квалификационная работа Суммаризация мультиязычных дискуссий в социальных сетях Уровень образования: магистратура …","url":["https://dspace.spbu.ru/bitstream/11701/32265/1/thesis.pdf"]}
{"year":"2021","title":"Multilingual ELMo and the Effects of Corpus Sampling","authors":["V Ravishankar, A Kutuzov, L Øvrelid, E Velldal - NoDaLiDa 2021, 2021"],"snippet":"… 3.2 Sampling Our initial starting point for collecting the language model training corpora were the CoNLL 2017 Wikipedia/Common Crawl dumps released as part of the shared task on Universal Dependencies parsing …","url":["https://ep.liu.se/ecp/178/ecp2021178.pdf#page=395"]}
{"year":"2021","title":"Multilingual Epidemic Event Extraction","authors":["S Mutuvi, E Boros, A Doucet, G Lejeune, A Jatowt… - International Conference on …, 2021"],"snippet":"Abstract In this paper, we focus on epidemic event extraction in multilingual and low-resource settings. The task of extracting epidemic events is defined as the detection of disease names and locations in a document. We experiment with a multilingual …","url":["https://link.springer.com/chapter/10.1007/978-3-030-91669-5_12"]}
{"year":"2021","title":"Multilingual Learning for Mild Cognitive Impairment Screening from a Clinical Speech Task","authors":["H Lindsay, P Müller, I Kröger, J Tröger, N Linz, A König… - 2021"],"snippet":"… al., 2020a). As semantic vectors are learned from large amounts of text corpora (usually Wikipedia and OSCAR common crawl), embedding quality heavily depends on the quantity of the available training data. While French …","url":["https://www.researchgate.net/profile/Hali-Lindsay-2/publication/354604884_Multilingual_Learning_for_Mild_Cognitive_Impairment_Screening_from_a_Clinical_Speech_Task/links/61420ec62db97e68051fbf26/Multilingual-Learning-for-Mild-Cognitive-Impairment-Screening-from-a-Clinical-Speech-Task.pdf"]}
{"year":"2021","title":"Multilingual Named Entity Recognition through Data and Model Transfer","authors":["S Palma-Suominen - 2021"],"snippet":"Page 1. UNIVERSITY OF HELSINKI FACULTY OF ARTS DEPARTMENT OF DIGITAL HUMANITIES Master's Thesis Multilingual Named Entity Recognition through Data and Model Transfer Saara Palma-Suominen 014291850 …","url":["https://helda.helsinki.fi/bitstream/handle/10138/330451/palma-suominen_saara_tutkielma_2021.pdf?sequence=2"]}
{"year":"2021","title":"Multilingual Neural Semantic Parsing for Low-Resourced Languages","authors":["M Xia, E Monti - arXiv preprint arXiv:2106.03469, 2021"],"snippet":"… It is trained for 100 languages on the CommonCrawl corpus, which is several orders of magnitude larger than the Wikipedia dump, especially for low-resourced languages. We use the public XLM-R Base model3 (12-layer, 768-hidden, 12 heads) in our experiment …","url":["https://arxiv.org/pdf/2106.03469"]}
{"year":"2021","title":"Multilingual Sentiment Analysis and Toxicity Detection for Text Messages in Russian","authors":["D Bogoradnikova, O Makhnytkina, A Matveev… - 2021 29th Conference of …, 2021"],"snippet":"… 179M pa- rameters, trained on cased texts in the top 104 languages with the largest Wikipedias), and xlm-roberta (270M parameters with 12-layers, 768-hidden-state, 3072 feed-forward hiddenstate, 8-heads, trained on 2.5 …","url":["https://ieeexplore.ieee.org/abstract/document/9435584/"]}
{"year":"2021","title":"Multilingual Transfer Learning for Code-Switched Language and Speech Neural Modeling","authors":["GI Winata - arXiv preprint arXiv:2104.06268, 2021"],"snippet":"Page 1. Multilingual Transfer Learning for Code-Switched Language and Speech Neural Modeling by Genta Indra Winata A Thesis Submitted to The Hong Kong University of Science and Technology in Partial Fulfillment of the Requirements for …","url":["https://arxiv.org/pdf/2104.06268"]}
{"year":"2021","title":"Multilingual Translation from Denoising Pre-Training","authors":["Y Tang, C Tran, X Li, PJ Chen, N Goyal, V Chaudhary…"],"snippet":"… al., 2019; Brown et al., 2020). Monolingual data is far more prevalent for low resource languages, particularly in resources such as Wikipedia or Commoncrawl, a version of the web. Recent work has explored monolingual denoising …","url":["https://aclanthology.org/2021.findings-acl.304.pdf"]}
{"year":"2021","title":"Multilingual Zero-Shot and Few-Shot Causality Detection","authors":["SM Reimann - 2021"],"snippet":"… BPE vocabulary with 100k subwords. Moreover, regarding training data, they build a CommonCrawl corpus for 100 languages, which is notably larger than the Wikipedia corpus used for mBERT. For pretraining XLM-R, they …","url":["https://www.diva-portal.org/smash/get/diva2:1569716/FULLTEXT01.pdf"]}
{"year":"2021","title":"Multimodal and Multilingual Embeddings for Large-Scale Speech Mining","authors":["PA Duquenne, H Gong, H Schwenk - Advances in Neural Information Processing …, 2021"],"snippet":"… Using a similarity metric in that multimodal embedding space, we perform mining of audio in German, French, Spanish and English from Librivox against billions of sentences from Common Crawl. This yielded more than twenty thousand hours of …","url":["https://proceedings.neurips.cc/paper/2021/file/8466f9ace6a9acbe71f75762ffc890f1-Paper.pdf"]}
{"year":"2021","title":"Multimodal datasets: misogyny, pornography, and malignant stereotypes","authors":["A Birhane, VU Prabhu, E Kahembwe - arXiv preprint arXiv:2110.01963, 2021"],"snippet":"… project to train a GDPR-Article(9)-Sensitive-URL classifier which revealed that, of the 1 Billion URLs they audited in the Common Crawl project, 155 million URLs fell into the sensitive category. The Realtoxicityprompts work …","url":["https://arxiv.org/pdf/2110.01963"]}
{"year":"2021","title":"Multimodal Emotion Recognition Using Transfer Learning on Audio and Text Data","authors":["JJ Deng, CHC Leung, Y Li - … Conference on Computational Science and Its …, 2021"],"snippet":"… As for text representation, we adopt fine-tuned Text-To-Text Transfer Transformer (T5) [23] model trained by a common crawl (C4) dataset to extract text embeddings. Operations of transfer learning have greatly accelerated model training and applied in specific domains …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86970-0_39"]}
{"year":"2021","title":"Multimodal MRI cerebral correlates of verbal fluency switching and its impairment in women with depression","authors":["L Domain, M Guillery, N Linz, A König, JM Batail… - NeuroImage: Clinical, 2021"],"snippet":"Background The search of biomarkers in the field of depression requires easy implementable tests that are biologically rooted. Qualitative analysis of verbal fluency tests (VFT) are good candidates, but its cerebral correlates are unknown …","url":["https://www.sciencedirect.com/science/article/pii/S2213158221003545"]}
{"year":"2021","title":"Multimodal or Text? Retrieval or BERT? Benchmarking Classifiers for the Shared Task on Hateful Memes","authors":["V Kougia, J Pavlopoulos"],"snippet":"… for English pre-trained on Common Crawl (Grave et al., 2018) (MNN:base). 2 The second variant em- ploys three BERT models, each fine-tuned on one of our tasks (see subsection 3.2), from which we extracted the CLS tokens …","url":["https://aclanthology.org/2021.woah-1.24.pdf"]}
{"year":"2021","title":"Multiplicative Position-aware Transformer Models for Language Understanding","authors":["Z Huang, D Liang, P Xu, B Xiang - arXiv preprint arXiv:2109.12788, 2021"],"snippet":"Page 1. Multiplicative Position-aware Transformer Models for Language Understanding Zhiheng Huang AWS AI zhiheng@amazon.com Davis Liang AWS AI liadavis@amazon.com Peng Xu AWS AI pengx@amazon.com …","url":["https://arxiv.org/pdf/2109.12788"]}
{"year":"2021","title":"Multitask Learning for Complaint Identification and Sentiment Analysis","authors":["A Singh, S Saha, M Hasanuzzaman, K Dey - Cognitive Computation, 2021"],"snippet":"In today's competitive business world, customer service is often at the heart of businesses that can help strengthen their brands. Resolution of cust.","url":["https://link.springer.com/article/10.1007/s12559-021-09844-7"]}
{"year":"2021","title":"MuRIL: Multilingual Representations for Indian Languages","authors":["S Khanuja, D Bansal, S Mehtani, S Khosla, A Dey… - arXiv preprint arXiv …, 2021"],"snippet":"… Monolingual Data: We collect monolingual data for the 17 languages mentioned above from the Common Crawl OSCAR corpus1 and Wikipedia2 … Wikipedia Common Crawl Wikipedia Common Crawl PMINDIA …","url":["https://arxiv.org/pdf/2103.10730"]}
{"year":"2021","title":"My House, My Rules: Learning Tidying Preferences with Graph Neural Networks","authors":["I Kapelyukh, E Johns - arXiv preprint arXiv:2111.03112, 2021"],"snippet":"… We load word embeddings from a FastText model [20], pre-trained on the Common Crawl dataset of over 600 billion tokens. Names of objects are provided to NeatNet either by the simulator or by an object detection system if deployed …","url":["https://arxiv.org/pdf/2111.03112"]}
{"year":"2021","title":"Nahuatl Neural Machine Translation Using Attention Based Architectures: A Comparative Analysis for RNNs and Transformers as a Mobile Application Service","authors":["SK Bello García, E Sánchez Lucero, E Bonilla Huerta… - … International Conference on …, 2021"],"snippet":"… Trained with 175 billions parameters and more than a trillion words from Common Crawl’s [27] and Wikipedia, making GPT-3 the largest language model at the time. The architecture is composed by 96 stacked encoder-decoder layers, 96 multi-head …","url":["https://link.springer.com/chapter/10.1007/978-3-030-89820-5_10"]}
{"year":"2021","title":"Named Entity Recognition and Classification on Historical Documents: A Survey","authors":["M Ehrmann, A Hamdi, EL Pontes, M Romanello… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Named Entity Recognition and Classification on Historical Documents: A Survey MAUD EHRMANN, Ecole Polytechnique Fédérale de Lausanne AHMED HAMDI, University of La Rochelle ELVYS LINHARES PONTES …","url":["https://arxiv.org/pdf/2109.11406"]}
{"year":"2021","title":"Named entity recognition and linking augmented with large-scale structured data","authors":["P Rychlikowski, A Lancucki, A Kaczmarek, B Najdecki… - Proceedings of the 8th …, 2021"],"snippet":"… possible. This line of reasoning applied especially to word-level embeddings. Ideally we wanted our systems to have, for every language, em- beddings trained on Wikipedia, Common Crawl1, and a collection of news articles. We …","url":["https://www.aclweb.org/anthology/2021.bsnlp-1.14.pdf"]}
{"year":"2021","title":"Named Entity Recognition and Normalization in Biomedical Literature: A Practical Case in SARS-CoV-2 Literature","authors":["ÁA Casero - 2021"],"snippet":"Page 1. Universidad Politécnica de Madrid Escuela Técnica Superior de Ingenieros Informáticos Máster Universitario en Inteligencia Artificial Trabajo Fin de Máster Named Entity Recognition and Normalization in Biomedical Literature: A …","url":["http://oa.upm.es/67933/1/TFM_ALVARO_ALONSO_CASERO.pdf"]}
{"year":"2021","title":"Naming unrelated words predicts creativity","authors":["JA Olson, J Nahas, D Chmoulevitch, SJ Cropper… - Proceedings of the National …, 2021"],"snippet":"Skip to main content. Main menu. Home; Articles: Current; Special Feature Articles - Most Recent; Special Features; Colloquia; Collected Articles; PNAS Classics; List of Issues. Front Matter: Front Matter Portal; Journal Club. News: For …","url":["https://www.pnas.org/content/pnas/118/25/e2022340118.full.pdf"]}
{"year":"2021","title":"Natural and artificial intelligence; natural and artificial language","authors":["D Santos - 2021"],"snippet":"… desinteligência, Inteligência, inteligência-do-cinema, inteligibilidade, inteligXancia, deligência, inteligencia, inteligência (0.85) 300 columns, obtained on 2 billion words, common crawl: fasttext, words: inteligencia (0.77), Inteligência, ainteligência …","url":["https://www.linguateca.pt/Diana/download/SantosSLATE2021.pdf"]}
{"year":"2021","title":"Natural language inference for Malayalam language using language agnostic sentence representation","authors":["S Renjit, S Idicula - PeerJ Computer Science, 2021"],"snippet":"… It is available in many languages, including Malayalam. fastText provided pretrained word vector representations trained on Common Crawl and Wikipedia data in 294 languages, including Malayalam, which we utilized to create sentence representations …","url":["https://peerj.com/articles/cs-508/"]}
{"year":"2021","title":"Negation typology and general representation models for cross-lingual zero-shot negation scope resolution in Russian, French, and Spanish.","authors":["A Shaitarova, F Rinaldi - Proceedings of the 2021 Conference of the North …, 2021"],"snippet":"… We build on this work and compare mBERT with a new multilingual general purpose representation model, XLM-R. Unlike mBERT, XLM-R was pre-trained on more than two terabytes of filtered data collected by CommonCrawl …","url":["https://www.aclweb.org/anthology/2021.naacl-srw.3.pdf"]}
{"year":"2021","title":"Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding","authors":["Z Wang, L Wang, T Wu, T Li, G Wu - arXiv preprint arXiv:2109.04872, 2021"],"snippet":"Page 1. Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding Zhenzhi Wang, Limin Wang*, Tao Wu, Tianhao Li, Gangshan Wu State Key Laboratory for Novel Software Technology, Nanjing University …","url":["https://arxiv.org/pdf/2109.04872"]}
{"year":"2021","title":"Network embeddings from distributional thesauri for improving static word representations","authors":["A Jana, S Haldar, P Goyal - Expert Systems with Applications, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0957417421012276"]}
{"year":"2021","title":"Neural approaches to discourse coherence: modeling, evaluation and application","authors":["Y Farag - 2021"],"snippet":"Page 1. Neural approaches to discourse coherence: modeling, evaluation and application Youmna Farag Murray Edwards College This dissertation is submitted in September 2020 for the degree of Doctor of Philosophy Page 2. Page 3. Declaration …","url":["https://www.repository.cam.ac.uk/bitstream/handle/1810/322567/thesis.pdf?sequence=3"]}
{"year":"2021","title":"Neural Machine Translation: A Review of Methods, Resources, and Tools","authors":["Z Tan, S Wang, Z Yang, G Chen, X Huang, M Sun… - arXiv preprint arXiv …, 2020"],"snippet":"Page 1. Neural Machine Translation: A Review of Methods, Resources, and Tools Zhixing Tana,c,d, Shuo Wanga,c,d, Zonghan Yanga,c,d, Gang Chena,c,d, Xuancheng Huanga,c,d, Maosong Suna,c,d,e, Yang Liua,b,c,d,e,∗ …","url":["https://arxiv.org/pdf/2012.15515"]}
{"year":"2021","title":"Neural Methods for Answer Passage Retrieval over Sparse Collections","authors":["D Cohen - 2021"],"snippet":"Page 1. University of Massachusetts Amherst ScholarWorks@UMass Amherst Doctoral Dissertations Dissertations and Theses April 2021 Neural Methods for Answer Passage Retrieval over Sparse Collections Daniel Cohen …","url":["https://scholarworks.umass.edu/cgi/viewcontent.cgi?article=3157&context=dissertations_2"]}
{"year":"2021","title":"Neural Representation Learning onTables","authors":["H Sun"],"snippet":"Page 1. From Tables to Knowledge (KDD'21): Jay Pujara, Pedro Szekely, Huan Sun, Muhao Chen Huan Sun The Ohio State University From Tables to Knowledge: Recent Advances in Table Understanding (Part III) Neural …","url":["https://pdfs.semanticscholar.org/62f0/dc599074928447c2582a136e08151255f79c.pdf"]}
{"year":"2021","title":"Neural Search: Learning Query and Product Representations in Fashion E-commerce","authors":["L Kumar, S Sarkar - arXiv preprint arXiv:2107.08291, 2021"],"snippet":"Page 1. Neural Search: Learning Query and Product Representations in Fashion E-commerce Sagnik Sarkar* sagnik.sarkar@myntra.com Myntra Designs Pvt. Ltd. India Lakshya Kumar* lakshya.kumar@myntra.com Myntra Designs Pvt. Ltd. India …","url":["https://arxiv.org/pdf/2107.08291"]}
{"year":"2021","title":"Neural Sequential Modeling and Applications","authors":["G Lai - 2021"],"snippet":"Page 1. Neural Sequential Modeling and Applications Guokun Lai CMU-LTI-21-010 Language Technology Institute School of Computer Science Carnegie Mellon University 5000 Forbes Ave., Pittsburgh, PA 15213 www.lti.cs.cmu.edu …","url":["https://www.lti.cs.cmu.edu/sites/default/files/lai%2C%20guokun%20-%20Thesis.pdf"]}
{"year":"2021","title":"Neural Supervised Domain Adaptation by Augmenting Pre-trained Models with Random Units","authors":["S Meftah, N Semmar, Y Tamaazousti, H Essafi, F Sadat - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Neural Supervised Domain Adaptation by Augmenting Pre-trained Models with Random Units Sara Meftah∗, Nasredine Semmar∗, Youssef Tamaazousti∗, Hassane Essafi∗, Fatiha Sadat+ ∗CEA-List, Université Paris …","url":["https://arxiv.org/pdf/2106.04935"]}
{"year":"2021","title":"Neural Text Generation with Artificial Negative Examples to Address Repeating and Dropping Errors","authors":["K Shirai, K Hashimoto, A Eriguchi, T Ninomiya, S Mori - Journal of Natural Language …, 2021"],"snippet":"… Both Japanese and English sentences were preprocessed as recommended in WAT'15.5 For the De-En task, we used the datasets provided by WMT'166 as well as all parallel corpora (Europarl v7 …","url":["https://www.jstage.jst.go.jp/article/jnlp/28/3/28_751/_pdf"]}
{"year":"2021","title":"Neural Text Generation with Artificial Negative Examples","authors":["K Shirai, K Hashimoto, A Eriguchi, T Ninomiya, S Mori - arXiv preprint arXiv …, 2020"],"snippet":"… For the De-En task, we used the datasets provided by WMT'166 and used all parallel corpora (Europarl v7, Common Crawl corpus, and News Commentary v11) for training, NewsTest2013 for development, and NewsTest2014 for testing …","url":["https://arxiv.org/pdf/2012.14124"]}
{"year":"2021","title":"Neural Transfer Learning for Domain Adaptation in Natural Language Processing","authors":["S Meftah - 2021"],"snippet":"Page 1. Neural Transfer Learning for Domain Adaptation in Natural Language Processing Apprentissage par Transfert Neuronal pour l'Adaptation aux Domaines en Traitement Automatique de la Langue Thèse de doctorat de l'université Paris-Saclay …","url":["https://www.theses.fr/2021UPASG021.pdf"]}
{"year":"2021","title":"Neural Transfer Learning with Transformers for Social Science Text Analysis","authors":["S Wankmüller - arXiv preprint arXiv:2102.02111, 2021"],"snippet":"Page 1. Neural Transfer Learning with Transformers for Social Science Text Analysis Sandra Wankmüller Ludwig-Maximilians-University Munich sandra.wankmueller@ gsi.lmu.de Abstract. During the last years, there have been …","url":["https://arxiv.org/pdf/2102.02111"]}
{"year":"2021","title":"NeurIPS 2020 EfficientQA Competition: Systems, Analyses and Lessons Learned","authors":["S Min, J Boyd-Graber, C Alberti, D Chen, E Choi… - arXiv preprint arXiv …, 2021"],"snippet":"… T5-XL+SSM T5 [13] is a text-to-text Transformer language model utilizing an encoder-decoder architecture that was pre-trained using a “span corruption” objective on a cleaned and de-duplicated subset of Common Crawl. Roberts et al …","url":["https://arxiv.org/pdf/2101.00133"]}
{"year":"2021","title":"New Synonyms Extraction Model Based on a Novel Terms Weighting Scheme","authors":["AH Ababneh, J Lu, Q Xu"],"snippet":"Page 1. 171 JIOS, VOL. 45. NO. 1 (2021), PP. 171-221 JIOS, VOL. 45, NO. 1 (2021) SUBMITTED 08/20; ACCEPTED 11/20 New Synonyms Extraction Model Based on a Novel Terms Weighting Scheme Ahmad Hussein Ababneh ahmad.ababneh@aum.edu.jo …","url":["https://jios.foi.hr/index.php/jios/article/view/1526/904"]}
{"year":"2021","title":"News Headline Summarization","authors":["A Kulkarni, A Shivananda, A Kulkarni - Natural Language Processing Projects, 2022"],"snippet":"… This transformer model is trained on the Colossal Clean Common Crawl (C4) data set. The architecture is integrated to devise text2text model. And hence T5 can also be used for many text-to-text-based applications. …","url":["https://link.springer.com/chapter/10.1007/978-1-4842-7386-9_10"]}
{"year":"2021","title":"NICT-2 Translation System at WAT-2021: Applying a Pretrained Multilingual Encoder-Decoder Model to Low-resource Language Pairs","authors":["K Imamura, E Sumita - Malay (Ms)"],"snippet":"… model can learn multiple languages. The published pretrained mBART model2 consists of a 12-layer encoder and decoder with a model dimension of 1,024 on 16 heads. This model was trained on 25 languages in the Common Crawl corpus (Wenzek et al., 2019) …","url":["https://aclanthology.org/2021.wat-1.8.pdf"]}
{"year":"2021","title":"NLI Data Sanity Check: Assessing the Effect of Data Corruption on Model Performance","authors":["A Talman, M Apidianaki, S Chatzikyriakidis… - arXiv preprint arXiv …, 2021"],"snippet":"… 4The contexts for R1 and R2 consist of sentences re- trieved from Wikipedia. In R3 the contexts are retrieved from Wikipedia, News (Common Crawl), fiction, The Children's Book Test (CBT), formal spoken text and …","url":["https://arxiv.org/pdf/2104.04751"]}
{"year":"2021","title":"NLM at TREC 2020 Health Misinformation and Deep Learning Tracks","authors":["Y Mrabet, M Sarrouti, AB Abacha, S Gayen, T Goodwin…"],"snippet":"… Figure 1 presents an overview of our data pipeline, approaches and workflow. 1 Common Crawl News: https://github.com/commoncrawl/news-crawl 2 https://github.com/optimaize/ language-detector Page 2. 2 Authors Suppressed Due to Excessive Length Fig …","url":["https://trec.nist.gov/pubs/trec29/papers/NLM.HM.DL.pdf"]}
{"year":"2021","title":"No News is Good News: A Critique of the One Billion Word Benchmark","authors":["H Ngo, JGM Araújo, J Hui, N Frosst - arXiv preprint arXiv:2110.12609, 2021"],"snippet":"… We train models solely on Common Crawl web scrapes partitioned by year, and demonstrate that they perform worse on this task over time due to distributional shift. Analysis of this corpus reveals that it contains several examples of harmful text, as …","url":["https://arxiv.org/pdf/2110.12609"]}
{"year":"2021","title":"Node. js based Document Store for Web Crawling","authors":["D Bui - 2021"],"snippet":"… and internet archiving such as the Common Crawl Project. While the Internet Archive focuses on preservation, Common Crawl focuses on making web crawled data publicly accessible for scientific analysis and studies [3]. Common Crawl …","url":["https://scholarworks.sjsu.edu/cgi/viewcontent.cgi?article=2039&context=etd_projects"]}
{"year":"2021","title":"NormFormer: Improved Transformer Pretraining with Extra Normalization","authors":["S Shleifer, J Weston, M Ott - arXiv preprint arXiv:2110.09456, 2021"],"snippet":"During pretraining, the Pre-LayerNorm transformer suffers from a gradient magnitude mismatch: gradients at early layers are much larger than at later layers. These issues can be alleviated by our proposed NormFormer architecture, which …","url":["https://arxiv.org/pdf/2110.09456"]}
{"year":"2021","title":"Not a Mirror, but an Engine: Digital Methods for Contextual Analysis of “Social Big Data”","authors":["JA Schwarz - Digital Human Sciences"],"snippet":"… In SNA, for example, the set of social relations is always at least two-dimensional 16 and is usually represented in the form of a node table (listing all the nodes 14 There are, eg, open science projects like …","url":["https://stockholmuniversitypress.se/site/books/10.16993/bbk/download/7869/#page=34"]}
{"year":"2021","title":"Not Far Away, Not So Close: Sample Efficient Nearest Neighbour Data Augmentation via MiniMax","authors":["E Kamalloo, M Rezagholizadeh, P Passban, A Ghodsi - arXiv preprint arXiv …, 2021"],"snippet":"… The aim of this stage is to find interpretable unannotated augmented samples that are semantically close to training data. For this purpose, we use one of the sentence repositories from SentAugment (Du et al., 2021), comprising …","url":["https://arxiv.org/pdf/2105.13608"]}
{"year":"2021","title":"Novelty Detection in Textual Data using Document Encoders and Self Attention.","authors":["A Parkhi - 2021"],"snippet":"… two-layer neural network. We try to minimize the negative log-likelihood of the next word for an entire corpus. GloVe vectors are trained on the common crawl dataset, consisting of 840B tokens and 2.2M different tokens. We use glove vectors with a spacy …","url":["https://repository.lib.ncsu.edu/bitstream/handle/1840.20/38854/etd.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"NPVec1: Word Embeddings for Nepali-Construction and Evaluation","authors":["P Koirala, NB Niraula"],"snippet":"… Common Crawl3 is a non-profit organization which collects data through web crawling and makes it publicly available. 3.1.3 Nepali Wikipedia Corpus We obtained Nepali Wikipedia corpus from Kaggle (Gaurav, 2020) …","url":["https://aclanthology.org/2021.repl4nlp-1.18.pdf"]}
{"year":"2021","title":"NUIG at TIAD 2021: Cross-lingual Word Embeddings for Translation Inference","authors":["S Banerjee, JP McCrae"],"snippet":"… concepts WordNet [13] graph analysis and cross-lingual word embeddings monolingual corpora of Common Crawl and Wikipedia [8] graph analysis relying on paths, synonyms, similarities and cardinality in the translation graph …","url":["https://sinaahmadi.github.io/docs/articles/ahmadi2021TIAD.pdf"]}
{"year":"2021","title":"NukeLM: Pre-Trained and Fine-Tuned Language Models for the Nuclear and Energy Domains","authors":["L Burke, K Pazdernik, D Fortin, B Wilson, R Goychayev… - arXiv preprint arXiv …, 2021"],"snippet":"… which tokens to predict in each batch rather than deciding offline, before training; it uses much larger batch sizes; it uses byte-level tokenization instead of character-level; and finally, it considers much more pre-training data …","url":["https://arxiv.org/pdf/2105.12192"]}
{"year":"2021","title":"Obtaining Better Static Word Embeddings Using Contextual Embedding Models","authors":["P Gupta, M Jaggi - arXiv preprint arXiv:2106.04302, 2021"],"snippet":"… We also evaluate some pre-trained 300 dimensional GLOVE (Pennington et al., 2014) and FASTTEXT (Bojanowski et al., 2016) models in Table 3. The GLOVE model was trained on Common-Crawl corpus of 840 Billion …","url":["https://arxiv.org/pdf/2106.04302"]}
{"year":"2021","title":"of black-box classifiers'","authors":["K Sinha - 2021"],"snippet":"… The embeddings used came from the 840b token Common Crawl pre-trained 300 dimensional GloVe vectors, limited to the vocabulary present in the SST dataset4. Ultimately, this model achieves 84% accuracy on the bi- nary SST classification task …","url":["https://www.researchgate.net/profile/Mario-Holubar/publication/352058956_Re_Replication_Study_of_'Generative_causal_explanations_of_black-box_classifiers'/links/60b77888a6fdcc476be1c863/Re-Replication-Study-of-Generative-causal-explanations-of-black-box-classifiers.pdf"]}
{"year":"2021","title":"Offensive Language Detection in Turkish Tweets with Bert Models","authors":["A Özberk, İ Çiçekli - 2021 6th International Conference on Computer …, 2021"],"snippet":"… XLM-RoBERTa that is a transformer-based cross-lingual model is pre-trained using a multilingual masNed language model objective on 2.5 TB of CommonCrawl data in 100 languages [15]. Our offensive language detection models use BERT …","url":["https://ieeexplore.ieee.org/abstract/document/9559000/"]}
{"year":"2021","title":"Offensive Language Identification in Low-resourced Code-mixed Dravidian languages using Pseudo-labeling","authors":["A Hande, K Puranik, K Yasaswini, R Priyadharshini… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Offensive Language Identification in Low-resourced Code-mixed Dravidian languages using Pseudo-labeling Adeep Handea,, Karthik Puranika,, Konthala Yasaswinia, Ruba Priyadharshinib, Sajeetha Thavareesanc …","url":["https://arxiv.org/pdf/2108.12177"]}
{"year":"2021","title":"Offensive, aggressive, and hate speech analysis: From data-centric to human-centered approach","authors":["J Kocoń, A Figas, M Gruza, D Puchalska… - Information Processing & …, 2021"],"url":["https://www.sciencedirect.com/science/article/pii/S0306457321001333"]}
{"year":"2021","title":"Omission of Information: Identifying Political Slant via an Analysis of Co-occurring Entities","authors":["J Ehrhardt, T Spinde, A Vardasbi, F Hamborg - 2021"],"snippet":"… 3 https://commoncrawl.org/ Page 6. Omission of Information: Identifying Political Slant via an Analysis of … 85 We scraped a total of 76.221 articles from nine different news outlets equally distributed over the three slant …","url":["https://epub.uni-regensburg.de/44939/1/isi_ehrhardt_et_al.pdf"]}
{"year":"2021","title":"OmniNet: Omnidirectional Representations from Transformers","authors":["Y Tay, M Dehghani, V Aribandi, J Gupta, P Pham, Z Qin… - arXiv preprint arXiv …, 2021"],"snippet":"… We run experiments on large-scale autoregressive (unidirectional) language modeling. We use two large-scale datasets, language modeling one billion (LM1B) (Chelba et al., 2013) and the Colossal Cleaned Common …","url":["https://arxiv.org/pdf/2103.01075"]}
{"year":"2021","title":"On a Utilitarian Approach to Privacy Preserving Text Generation","authors":["Z Xu, A Aggarwal, O Feyisetan, N Teissier - arXiv preprint arXiv:2104.11838, 2021"],"snippet":"… Due to the high dimensional nature of textual tasks and very large vocabulary sizes (eg 2.2M words for GLOVE common crawl (Pennington et al., 2014)), this can lead to adding a lot of noise for achieving the desired privacy …","url":["https://arxiv.org/pdf/2104.11838"]}
{"year":"2021","title":"On Anytime Learning at Macroscale","authors":["L Caccia, J Xu, M Ott, MA Ranzato, L Denoyer - arXiv preprint arXiv:2106.09563, 2021"],"snippet":"… For instance, every once in a while new data from common crawl is gathered to train even larger language models, or every few weeks new images and associated labels are collected to train better visual object recognizers …","url":["https://arxiv.org/pdf/2106.09563"]}
{"year":"2021","title":"On the dangers of stochastic parrots: Can language models be too big","authors":["EM Bender, T Gebru, A McMillan-Major, S Shmitchell - Proceedings of FAccT, 2021"],"snippet":"… show that models like GPT-3 trained with at least 570GB of data derived mostly from Common Crawl16 can generate sentences with high toxicity … 16https://commoncrawl.org/the-data/ 17GPT-3's training data is not openly …","url":["http://faculty.washington.edu/ebender/papers/Stochastic_Parrots.pdf"]}
{"year":"2021","title":"On the Language Coverage Bias for Neural Machine Translation","authors":["S Wang, Z Tu, Z Tan, S Shi, M Sun, Y Liu - arXiv preprint arXiv:2106.03297, 2021"],"snippet":"Page 1. On the Language Coverage Bias for Neural Machine Translation Shuo Wang1 Zhaopeng Tu2 Zhixing Tan1 Shuming Shi2 Maosong Sun1,3 Yang Liu1,3,4 1Dept. of Comp. Sci. & Tech., Institute for AI, BNRist Center, Tsinghua …","url":["https://arxiv.org/pdf/2106.03297"]}
{"year":"2021","title":"On the Learning Dynamics of Semi-Supervised Training for ASR}}","authors":["E Wallington, B Kershenbaum, O Klejch, P Bell - Proc. Interspeech 2021, 2021"],"snippet":"… Following initial analysis, we use these same varying-quality models to investigate iSST. 2.1. Language Models The LM training data consists of the Tagalog CommonCrawl data-set1 and four smaller Tagalog sets provided …","url":["https://www.isca-speech.org/archive/pdfs/interspeech_2021/wallington21_interspeech.pdf"]}
{"year":"2021","title":"On the Multilingual Capabilities of Very Large-Scale English Language Models","authors":["J Armengol-Estapé, OG Bonet, M Melero - arXiv preprint arXiv:2108.13349, 2021"],"snippet":"… models (Nozza et al., 2020). Even the multilingual ones2 such as mBERT (Devlin et al., 2018) or XLM-R (Conneau et al., 2019) employ large multilingual datasets based on Wikipedia or CommonCrawl. A very recent work trained …","url":["https://arxiv.org/pdf/2108.13349"]}
{"year":"2021","title":"On the Robustness of Neural Network: Attacks and Defenses","authors":["M Cheng - 2021"],"snippet":"Page 1. UCLA UCLA Electronic Theses and Dissertations Title On the Robustness of Neural Network: Attacks and Defenses Permalink https://escholarship.org/uc/item/3k2780bg Author Cheng, Minhao Publication …","url":["https://escholarship.org/content/qt3k2780bg/qt3k2780bg.pdf"]}
{"year":"2021","title":"On the Usability of Transformers-based models for a French Question-Answering task","authors":["O Cattan, C Servan, S Rosset - Recent Advances in Natural Language Processing …, 2021"],"snippet":"… Transformer-based language models such as BERT (Devlin et al., 2019) are pre-trained on largescale data collections sourced from Wikipedia or Common Crawl (CC) with one or multiple training objectives (masked …","url":["https://hal.archives-ouvertes.fr/hal-03336060/document"]}
{"year":"2021","title":"On the validity of pre-trained transformers for natural language processing in the software engineering domain","authors":["J von der Mosel, A Trautsch, S Herbold - arXiv preprint arXiv:2109.04738, 2021"],"snippet":"… Therefore, most of the available models are pre-trained by large companies on vast amounts of general domain data such as the entire English Wikipedia or the Common Crawl News dataset [5]. While these models achieve …","url":["https://arxiv.org/pdf/2109.04738"]}
{"year":"2021","title":"On-Line Retrieval of Health Information Based on Language Complexity, Information Customization and Information Quality","authors":["M Alfano - … and Communication Technologies for Ageing Well and …, 2021"],"snippet":"… initiative. The Web Data Commons (WDC)[30] contain all Microformat, Microdata and RDFa (Resource Description Framework in Attributes) data extracted from the open repository of Web crawl data named Common Crawl (CC) …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=9kQgEAAAQBAJ&oi=fnd&pg=PA1&dq=commoncrawl&ots=V3u-4Vufj1&sig=xQGsuQdnwOKvDO0cKw5-XZNUvjM"]}
{"year":"2021","title":"One to rule them all: Towards Joint Indic Language Hate Speech Detection","authors":["M Bhatia, TS Bhotia, A Agarwal, P Ramesh, S Gupta… - arXiv preprint arXiv …, 2021"],"snippet":"… A summary for each model is as follows: • XLM-RoBERTa: The pre-training of XLM-RoBERTa is based on 100 languages, using around 2.5TB of preprocessed CommonCrawl dataset to train cross-language …","url":["https://arxiv.org/pdf/2109.13711"]}
{"year":"2021","title":"Open Dataset Archive","authors":["T Weber"],"snippet":"… data: to name an example, while large corpora of tabular data from Web tables have been made available via CommonCrawl [5], the … Common Crawl The Common Crawl Foundation is a also a non-profit organization which …","url":["https://aic.ai.wu.ac.at/~polleres/supervised_theses/Thomas_Weber_BSc_2020.pdf"]}
{"year":"2021","title":"Open-Domain Conversational Search Assistant with Transformers","authors":["R Ferreira, M Leite, D Semedo, J Magalhaes - arXiv preprint arXiv:2101.08197, 2021"],"snippet":"… Page 6. 6 R. Ferreira et al. – Text-to-Text Transfer Transformer (T5) [26] is a text-to-text model based on the encoder-decoder Transformer architecture, pre-trained on the large C4 corpus, which was derived from Common …","url":["https://arxiv.org/pdf/2101.08197"]}
{"year":"2021","title":"Opening up Military Innovation: Causal Effects of 'Bottom-Up'Reforms to US Defense Research","authors":["ST Howell, J Rathje, J Van Reenen, J Wong - Available at SSRN 3825034, 2021"],"snippet":"Page 1. Opening up Military Innovation: Causal Effects of 'Bottom-Up' Reforms to US Defense Research Sabrina T. Howell, Jason Rathje, John Van Reenen and Jun Wong ∗ April 8, 2021 Abstract When investing in research …","url":["https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3825034"]}
{"year":"2021","title":"OpenViDial: A Large-Scale, Open-Domain Dialogue Dataset with Visual Contexts","authors":["Y Meng, S Wang, Q Han, X Sun, F Wu, R Yan, J Li - arXiv preprint arXiv:2012.15015, 2020"],"snippet":"… and the open-sourced OCR model suffices to fulfill our need. With text-free images in hand, we pair them with texts. Texts are randomly selected from the CommonCrawl English corpus, then added to the images. Texts in images …","url":["https://arxiv.org/pdf/2012.15015"]}
{"year":"2021","title":"Operationalizing a National Digital Library: The Case for a Norwegian Transformer Model","authors":["PE Kummervold, J De la Rosa, F Wetjen, SA Brygfjeld - arXiv preprint arXiv …, 2021"],"snippet":"… For resource-rich languages such as Chinese, En- glish, French, and Spanish, collections of texts from open sources such as Wikipedia (2021a), variations of Common Crawl data (2021), and other open-source corpora such as the BooksCorpus (Zhu et al., 2015) are …","url":["https://arxiv.org/pdf/2104.09617"]}
{"year":"2021","title":"Opinion Classification via Word and Emoji Embedding Models with LSTM","authors":["I Rabbimov, S Kobilov, I Mporas - International Conference on Speech and Computer, 2021"],"snippet":"… Twitter. 1,661. 300. The UzbekFastText1 word vector is the fastText model trained with crawled texts published on Wikipedia using Common Crawl [36]. UzbekFastText2 was presented in [37] and it is a 300-dimensional Uzbek …","url":["https://www.researchgate.net/profile/Ilyos-Rabbimov/publication/354758888_Opinion_Classification_via_Word_and_Emoji_Embedding_Models_with_LSTM/links/614e7a91154b3227a8a8b8b8/Opinion-Classification-via-Word-and-Emoji-Embedding-Models-with-LSTM.pdf"]}
{"year":"2021","title":"Out-of-vocabulary but not meaningless: Evidence for semantic-priming effects in pseudoword processing","authors":["D Gatti, M Marelli, L Rinaldi - 2021"],"snippet":"… The model was trained on Common Crawl (around 630 billion words) and Wikipedia (around 9 billion words) using the Continuous Bag of … The model was trained on Common Crawl and Italian Wikipedia (around 11 billion words) using …","url":["https://psyarxiv.com/y2fm3/download?format=pdf"]}
{"year":"2021","title":"Overcoming Data Challenges in Machine Translation","authors":["H Khayrallah - 2021"],"snippet":"Page 1. OVERCOMING DATA CHALLENGES IN MACHINE TRANSLATION by Huda Khayrallah A dissertation submitted to The Johns Hopkins University in conformity with the requirements for the degree of Doctor of Philosophy. Baltimore, Maryland July, 2021 …","url":["https://jscholarship.library.jhu.edu/bitstream/handle/1774.2/64381/KHAYRALLAH-DISSERTATION-2021.pdf?sequence=1"]}
{"year":"2021","title":"Overlapping Community Detection in Social Networks","authors":["A Panchal - 2021"],"snippet":"Page 1. San Jose State University SJSU ScholarWorks Master's Projects Master's Theses and Graduate Research Spring 5-25-2021 Overlapping Community Detection in Social Networks Akshar Panchal Follow this and additional …","url":["https://scholarworks.sjsu.edu/cgi/viewcontent.cgi?article=2011&context=etd_projects"]}
{"year":"2021","title":"Overview of Touché 2021: Argument Retrieval","authors":["A Bondarenko, L Gienapp, M Fröbe, M Beloucif… - 2021"],"snippet":"… Other argument retrieval systems such as ArgumenText [3] and TARGER [4] take advantage of the large web document collection Common Crawl, but their ability to reliably retrieve arguments to support sides in a decision process is limited …","url":["https://webis.de/downloads/publications/papers/bondarenko_2021b.pdf"]}
{"year":"2021","title":"PAGnol: An Extra-Large French Generative Model","authors":["J Launay, EL Tommasone, B Pannier, F Boniface… - arXiv preprint arXiv …, 2021"],"snippet":"… Having been pre-trained on Common Crawl-based corpora, our models are certainly not immune from toxic content generation. Gehman et al. (2020) explored two main ways to filter toxic outputs, one being based on the pursuing the pretraining on less …","url":["https://arxiv.org/pdf/2110.08554"]}
{"year":"2021","title":"PALI at SemEval-2021 Task 2: Fine-Tune XLM-RoBERTa for Word in Context Disambiguation","authors":["S Xie, J Ma, H Yang, L Jiang, Y Mo, J Shen - arXiv preprint arXiv:2104.10375, 2021"],"snippet":"… 2017). Especially, the XLM-RoBERTa (Conneau et al., 2020) model is a newly released large cross-lingual language model based on RoBERTa and is trained on 2.5TB filtered CommonCrawl data in 100 languages. Different …","url":["https://arxiv.org/pdf/2104.10375"]}
{"year":"2021","title":"PAN: Prototype-based Adaptive Network for Robust Cross-modal Retrieval","authors":["Z Zeng, S Wang, N Xu, W Mao - Proceedings of the 44th International ACM SIGIR …, 2021"],"snippet":"Page 1. PAN: Prototype-based Adaptive Network for Robust Cross-modal Retrieval Zhixiong Zeng, Shuai Wang, Nan Xu, Wenji Mao∗ SKL-MCCS, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China …","url":["https://dl.acm.org/doi/abs/10.1145/3404835.3462867"]}
{"year":"2021","title":"PanGu-$\\alpha $: Large-scale Autoregressive Pretrained Chinese Language Models with Auto-parallel Computation","authors":["W Zeng, X Ren, T Su, H Wang, Y Liao, Z Wang, X Jiang… - arXiv preprint arXiv …, 2021"],"snippet":"… To build Chinese corpus with comprehensive coverage, we collect a large amount of data from a wide range of resources, including Common Crawl, e-Books, encyclopedias, news, and so on … Common Crawl 176.2B 62.81 …","url":["https://arxiv.org/pdf/2104.12369"]}
{"year":"2021","title":"PARADISE: Exploiting Parallel Data for Multilingual Sequence-to-Sequence Pretraining","authors":["M Reid, M Artetxe - arXiv preprint arXiv:2108.01887, 2021"],"snippet":"Page 1. PARADISE: Exploiting Parallel Data for Multilingual Sequence-to-Sequence Pretraining Machel Reid The University of Tokyo machelreid@weblab.tu-tokyo.ac.jp Mikel Artetxe Facebook AI Research artetxe@fb.com Abstract …","url":["https://arxiv.org/pdf/2108.01887"]}
{"year":"2021","title":"Parallel feature weight decay algorithms for fast development of machine translation models","authors":["E Biçici - Machine Translation, 2021"],"snippet":"Parallel feature weight decay algorithms, parfwd, are engineered for languageand task-adaptive instance selection to build distinct machine translation (.","url":["https://link.springer.com/article/10.1007/s10590-021-09275-z"]}
{"year":"2021","title":"Parallelizing Legendre Memory Unit Training","authors":["N Chilkuri, C Eliasmith - arXiv preprint arXiv:2102.11417, 2021"],"snippet":"… After showing that the training of this simplified variant can be 1https:// commoncrawl.org … Therefore, we construct parameter efficient models that em- ploy just the DN layer, with d = 1 and θ = maxlen. We use 300D Glove …","url":["https://arxiv.org/pdf/2102.11417"]}
{"year":"2021","title":"Paraphrase Generation as Unsupervised Machine Translation","authors":["C Fan, Y Tian, Y Meng, N Peng, X Sun, F Wu, J Li - arXiv preprint arXiv:2109.02950, 2021"],"snippet":"… LDA and K-means algorithms are performed on part of the the CommonCrawl corpus containing 10 billion English tokens … An overview of deriving the UMT-Multi model is shown in Figure 1. Up to now, UMT-Multi is purely based on unlabeled commoncrawl corpus …","url":["https://arxiv.org/pdf/2109.02950"]}
{"year":"2021","title":"Pathogens Are Linked to Human Moral Systems Across Time and Space","authors":["M Atari, NK Reimer, J Graham, J Hoover, B Kennedy… - 2021"],"snippet":"Infectious diseases have been an impending threat to the survival of individuals and groups throughout our evolutionary history. As a result, humans have developed psychological pathogen-avoidance mechanisms and groups have developed …","url":["https://psyarxiv.com/tnyh9/download?format=pdf"]}
{"year":"2021","title":"Pegasus@ Dravidian-CodeMix-HASOC2021: Analyzing Social Media Content for Detection of Offensive Text","authors":["PK Jada, K Yasaswini, K Puranik, A Sampath… - arXiv preprint arXiv …, 2021"],"snippet":"… One of the reason for the unparalleled performance is that it was trained on a mammoth 2.5 TB of CommonCrawl data [25]. It was trained with MLM loss as it’s objective on 100 different languages, and it shares similar training routine as the …","url":["https://arxiv.org/pdf/2111.09836"]}
{"year":"2021","title":"PENGEMBANGAN APLIKASI INFORMATION GATHERING MENGGUNAKAN METODE HYBRID SCAN BERBASIS GRAPHICAL USER INTERFACE","authors":["M Riasetiawan, A Wisnuaji, D Hariyadi, T Febrianto - Cyber Security dan Forensik …, 2021"],"snippet":"… F-Secure Riddler, Cert Spotter, Hacker Target, Certificate Search, DN Dumpster, AlienFault, Rapid DNS, URL Scan, Threat Crowd, Wayback Machine, dan Common Crawl … q=\"+self.url, 1 2 \"https://riddler.io/search …","url":["http://202.0.92.5/saintek/cybersecurity/article/download/2449/1994"]}
{"year":"2021","title":"Penn-Helsinki Parsed Corpus of Early Modern English: First Parsing Results and Analysis","authors":["S Kulick, N Ryant, B Santorini - arXiv preprint arXiv:2112.08532, 2021"],"snippet":"We present the first parsing results on the Penn-Helsinki Parsed Corpus of Early Modern English (PPCEME), a 1.9 million word treebank that is an important resource for research in syntactic change. We describe key features of PPCEME that …","url":["https://arxiv.org/pdf/2112.08532"]}
{"year":"2021","title":"Personalized sentiment classification of customer reviews via an interactive attributes attention model","authors":["Y Zhang, J Wang, X Zhang - Knowledge-Based Systems, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0950705121003981"]}
{"year":"2021","title":"Phishing Detection Methods: A Review","authors":["M Alanezi - 2021"],"snippet":"Abstract The web has turned into a principal part of our conventional social and financial activities. The web isn't significant for singular clients just yet additionally for associations, since associations that offer web-based exchanging can accomplish …","url":["https://techniumscience.com/index.php/technium/article/view/4973/1723"]}
{"year":"2021","title":"PhoMT: A High-Quality and Large-Scale Benchmark Dataset for Vietnamese-English Machine Translation","authors":["L Doan, LT Nguyen, NL Tran, T Hoang, DQ Nguyen - arXiv preprint arXiv:2110.12199, 2021"],"snippet":"We introduce a high-quality and large-scale Vietnamese-English parallel dataset of 3.02M sentence pairs, which is 2.9M pairs larger than the benchmark Vietnamese-English machine translation corpus IWSLT15. We conduct experiments comparing strong …","url":["https://arxiv.org/pdf/2110.12199"]}
{"year":"2021","title":"Planning with Entity Chains for Abstractive Summarization","authors":["S Narayan, Y Zhao, J Maynez, G Simoes, R McDonald - arXiv preprint arXiv …, 2021"],"snippet":"… Pre-training Datasets. Following Zhang et al. (2019), our model pretraining also relied on two large web corpus which were processed to look like plain text: (i) C4 (Raffel et al., 2019) is composed of 350M Web-pages …","url":["https://arxiv.org/pdf/2104.07606"]}
{"year":"2021","title":"Planning with Learned Entity Prompts for Abstractive Summarization","authors":["S Narayan, Y Zhao, J Maynez, G Simões, V Nikolaev… - Transactions of the …, 2021"],"snippet":"We introduce a simple but flexible mechanism to learn an intermediate plan to ground the generation of abstractive summaries. Specifically, we prepend (or prompt) target summaries with entity chains—ordered sequences of entities mentioned in the …","url":["https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00438/108867"]}
{"year":"2021","title":"Polish Web resources described in the\" Polish World\" directory (1997). Characteristics of domains and their conservation state","authors":["M Wilkowski - Archiwa-Kancelarie-Zbiory, 2020"],"snippet":"… Sep. 2008, p. 6. https://hal-bnf.archives-ouvertes.fr/hal-01098538 (accessed 19.09.2020). 18 Common Crawl, https://commoncrawl.org/ (accessed 12.09.2020). 19 HTTP Archive, https://httparchive.org/ (accessed 12.09.2020). Page 8. 126 Marcin Wilkowski …","url":["https://apcz.umk.pl/czasopisma/index.php/AKZ/article/download/AKZ.2020.005/28213"]}
{"year":"2021","title":"Political Attacks in 280 Characters or Less: A New Tool for the Automated Classification of Campaign Negativity on Social Media","authors":["V Petkevic, A Nai - American Politics Research, 2021"],"snippet":"Negativity in election campaign matters. To what extent can the content of social media posts provide a reliable indicator of candidates' campaign negativity? We introduce and critically assess an automated classification procedure that we trained …","url":["https://journals.sagepub.com/doi/pdf/10.1177/1532673X211055676"]}
{"year":"2021","title":"Political Polarization in Online News Consumption","authors":["K Garimella, T Smith, R Weiss, R West - arXiv preprint arXiv:2104.06481, 2021"],"snippet":"… The Common Crawl project1 regularly crawls the Web and releases datasets of the discovered webpages on a regular schedule, which are generally considered the most extensive publicly available sources …","url":["https://arxiv.org/pdf/2104.06481"]}
{"year":"2021","title":"PoolRank: Max/Min Pooling-based Ranking Loss for Listwise Learning & Ranking Balance","authors":["Z Chen, C Eickhoff - arXiv preprint arXiv:2108.03586, 2021"],"snippet":"… Our GloVe word embeddings are trained on the Common Crawl dataset 2. We use an Adam optimizer [13] with 0.0001 learning rate to train all models … Most queries 1https://www.tensor ow.org/ 2https://commoncrawl …","url":["https://arxiv.org/pdf/2108.03586"]}
{"year":"2021","title":"Postal address extraction from the web: a comprehensive survey","authors":["M Kayed, S Dakrory, AA Ali - Artificial Intelligence Review, 2021"],"snippet":"The Web is a source of information for Location-Based Service (LBS) applications. These applications lack postal addresses for the user's Point of In.","url":["https://link.springer.com/article/10.1007/s10462-021-09983-1"]}
{"year":"2021","title":"Practical Wavelet Tree Construction","authors":["P Dinklage, J Ellert, J Fischer, F Kurpicz, M Löbel - Journal of Experimental …, 2021","VIN LÖBEL - 2021"],"snippet":"… alphabet. We denote the resulting file by RuWB. Suffix Array of Common Crawl We construct the suffix array [35] of CommonCrawl, which is a permutation of the positive integers smaller than the length of CommonCrawl. We …","url":["https://dl.acm.org/doi/pdf/10.1145/3457197","https://www.kurpicz.org/assets/pdfs/jea_2021.pdf"]}
{"year":"2021","title":"Pre-trained Language Models for Tagalog with Multi-source Data","authors":["S Jiang, Y Fu, X Lin, N Lin - CCF International Conference on Natural Language …, 2021"],"snippet":"… [19] to process, filter and classify Common Crawl, which is a non-profit organization that produces and maintains an open, freely available repository of crawled data from the web. The filtering step used to create OSCAR involves …","url":["https://link.springer.com/chapter/10.1007/978-3-030-88480-2_17"]}
{"year":"2021","title":"Pre-trained Language Models in Biomedical Domain: A Survey from Multiscale Perspective","authors":["B Wang, Q Xie, J Pei, P Tiwari, Z Li - arXiv preprint arXiv:2110.05006, 2021"],"snippet":"Pre-trained language models have been the de facto paradigm for most natural language processing (NLP) tasks. In the biomedical domain, which also benefits from NLP techniques, various pre-trained language models were proposed by …","url":["https://arxiv.org/pdf/2110.05006"]}
{"year":"2021","title":"Pre-trained Models and Evaluation Data for the Myanmar Language","authors":["S Jiang, X Huang, X Cai, N Lin - International Conference on Neural Information …, 2021"],"snippet":"… On the one hand, we utilized all the Myanmar data from the OSCAR corpus 3 , a humongous multilingual corpus whose texts all come from the Common Crawl corpus 4 . In addition, articles on CC-100 [16] were also used as part of our corpus …","url":["https://link.springer.com/chapter/10.1007/978-3-030-92310-5_52"]}
{"year":"2021","title":"Pre-Trained Transformer-Based Language Models for Sundanese","authors":["W Wongso, H Lucky, D Suhartono - 2021"],"snippet":"… Data The dataset used during pre-training consists of Sundanese subsets of multilingual common-crawl corpora of OSCAR [18], CC-100 [19, 20], and C4 [21]. They are of the size 141KB, 15MB, and 464MB, respectively. Seeing …","url":["https://www.researchsquare.com/article/rs-907893/latest.pdf"]}
{"year":"2021","title":"Pre-trained Word Embeddings for Malayalam Language: A Review","authors":["KR Rahmath, PCR Raj, PC Rafeeque - … on Artificial Intelligence and Smart Systems …, 2021"],"snippet":"… pre-trained on 17 In- dian languages, and their transliterated counterparts including Malayalam [8]. The corpora used for pre-training process are Wikipedia, Common Crawl, PMINDIA and … This dataset has been augmented …","url":["https://ieeexplore.ieee.org/abstract/document/9396042/"]}
{"year":"2021","title":"Pre-training Methods in Information Retrieval","authors":["Y Fan, X Xie, Y Cai, J Chen, X Ma, X Li, R Zhang, J Guo… - arXiv preprint arXiv …, 2021"],"snippet":"The core of information retrieval (IR) is to identify relevant information from large-scale resources and return it as a ranked list to respond to user's information need. Recently, the resurgence of deep learning has greatly advanced this field and leads …","url":["https://arxiv.org/pdf/2111.13853"]}
{"year":"2021","title":"Predicting cross-linguistic adjective order with information gain","authors":["W Dyer, R Futrell, Z Liu, G Scontras - arXiv preprint arXiv:2012.15263, 2020"],"snippet":"… 4.1 Data Our study relies on two types of source data, both extracted from the CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal De- pendencies (Ginter et al., 2017; Zeman et al., 2017) a set of …","url":["https://arxiv.org/pdf/2012.15263"]}
{"year":"2021","title":"Predicting Depression and Suicide Ideation in the Canadian Population Using Social Media Data","authors":["R Skaik - 2021"],"snippet":"Page 1. Predicting Depression and Suicide Ideation in the Canadian Population Using Social Media Data by Ruba Skaik Thesis submitted to the University of Ottawa in partial Fulfillment of the requirements for the Ph.D. degree in Computer Science …","url":["https://ruor.uottawa.ca/bitstream/10393/42346/7/Skaik_Ruba_2021_thesis.pdf"]}
{"year":"2021","title":"Predicting leadership perception with large-scale natural language data","authors":["S Bhatia, CY Olivola, N Bhatia, A Ameen - The Leadership Quarterly, 2021"],"snippet":"","url":["https://www.sciencedirect.com/science/article/pii/S1048984321000400"]}
{"year":"2021","title":"Predicting Next Dialogue Action in Emotionally Loaded Conversation","authors":["D Deksne, R Skadiņš - Proceedings of the Future Technologies Conference, 2021"],"snippet":"… the Latvian Wikipedia data 2 and trained with our internal monolingual corpora, as well as several BERT models: multilingual BERT [26], BERT-LV [27], Language-agnostic BERT Sentence Embedding (LaBSE) [28], and the multilingual XLM-R model trained …","url":["https://link.springer.com/chapter/10.1007/978-3-030-89906-6_19"]}
{"year":"2021","title":"Predicting the Performance of Multilingual NLP Models","authors":["A Srinivasan, S Sitaram, T Ganu, S Dandapat, K Bali… - arXiv preprint arXiv …, 2021"],"snippet":"… The pretraining procedure makes use of only unlabelled data (raw text), which is easy to obtain for a large number of languages from various sources including CommonCrawl1 and Wikipedia dumps2. These models can then be used to solve a …","url":["https://arxiv.org/pdf/2110.08875"]}
{"year":"2021","title":"Preprint from: https://www. gipp. com/pub","authors":["T Spinde, L Rudnitckaia, K Sinha, F Hamborg"],"snippet":"… We scraped articles from both news outlets, published in the last decade, from 2010 to 2020, from Common Crawl [4]. For preprocessing, we use Genism simple preprocessing and generate n-grams. 3.4 Linear mapping between vector spaces …","url":["https://www.gipp.com/wp-content/papercite-data/pdf/spinde2021.pdf"]}
{"year":"2021","title":"PREPRINT–ARTICLE IN PRESS: Kennedy, B., Atari, M., Mostafazadeh Davani, A., Hoover, J., Omrani, A., Graham, J., & Dehghani, M.(in press). Moral Concerns are …","authors":["B Kennedy"],"snippet":"… of posts, GloVe word embedding vectors (Pennington et al., 2014), which were trained on text from the Common Crawl4, were mapped to the individual words in the posts and subsequently averaged element-wise (ie, across the dimensions of the embedding). The …","url":["https://files.osf.io/v1/resources/uqmty/providers/osfstorage/5eb44939a2199500cd5d4b7c?format=pdf&action=download&direct&version=2"]}
{"year":"2021","title":"Prescriptions in NLP with Transformer Models","authors":["OEC Zaragoza"],"snippet":"… is optimized BERT. XLMR is trained with a filtered version of CommonCrawl data with 100 languages and it uses a subword tokenization, SentencePiece [20] with a unigram language model [19]. mT5 Multilingual TexttoText …","url":["https://payberah.github.io/files/download/students/omar_contreras_master_thesis.pdf"]}
{"year":"2021","title":"Pretrained models and evaluation data for the Khmer language","authors":["S Jiang, S Fu, N Lin, Y Fu - Tsinghua Science and Technology, 2021"],"snippet":"… On the one hand, we use all the central Khmer data from the OSCAR corpus , a large multilingual corpus whose texts come from the Common Crawl corpus . Although Common Crawl comprises scraped data from the Internet and covers a …","url":["https://ieeexplore.ieee.org/iel7/5971803/9645434/09645441.pdf"]}
{"year":"2021","title":"Prevalence in News Media of Two Competing Hypotheses about COVID-19 Origins","authors":["D Rozado - Social Sciences, 2021"],"snippet":"… 2. Methods. The textual content of news and opinion articles from the outlets listed in Figure 1 is available in the outlets online domains and/or public cache repositories such as Google cache, The Internet Wayback Machine …","url":["https://www.mdpi.com/2076-0760/10/9/320/htm"]}
{"year":"2021","title":"Prevalence of Prejudice-Denoting Words in News Media Discourse: A Chronological Analysis","authors":["D Rozado, M Al-Gharbi, J Halberstadt - Social Science Computer Review, 2021"],"snippet":"This work analyzes the prevalence of words denoting prejudice in 27 million news and opinion articles written between 1970 and 2019 and published in 47 of the most popular news media outlets in the...","url":["https://journals.sagepub.com/doi/abs/10.1177/08944393211031452"]}
{"year":"2021","title":"Preventing author profiling through zero-shot multilingual back-translation","authors":["DI Adelani, M Zhang, X Shen, A Davody, T Kleinbauer… - arXiv preprint arXiv …, 2021"],"snippet":"… The languages chosen are German (DE), Spanish (ES), French (FR), Japanese (JA), Russian (RU), and Chinese (ZH) based on the large amount of resources they have on OPUS (Tiedemann, 2012) and Common Crawl corpora 2 …","url":["https://arxiv.org/pdf/2109.09133"]}
{"year":"2021","title":"Price Determinants of Airbnb Apartments","authors":["A Piispanen - 2021"],"snippet":"Page 1. Antti Piispanen Price Determinants of Airbnb Apartments An Approach with Deep Language Representations Master of Science Thesis Faculty of Information Technology and Communication Sciences Examiner …","url":["https://trepo.tuni.fi/bitstream/handle/10024/130318/PiispanenAntti.pdf?sequence=2"]}
{"year":"2021","title":"Prioritizing Original News on Facebook","authors":["X Ni, S Bu, I Markov - arXiv preprint arXiv:2102.08465, 2021"],"snippet":"… There are other open-source crawlers that serve the same purpose. Common Crawl3 is a well-maintained open repository of Web crawl data that can be accessed and analyzed by anyone. We limit news articles in the graph to …","url":["https://arxiv.org/pdf/2102.08465"]}
{"year":"2021","title":"Privacy and AI Ethics–Understanding the convergences and tensions for the responsible development of machine learning","authors":["I Cofone, O Abimana, B Bonan, E Grand-Pierre, A Qarri"],"snippet":"Context. The democratization of mobile systems and the development of information technologies have been accompanied by a massive increase in the amount and the diversity of data collected about individuals, often referred to as Big Data …","url":["https://sebastiengambs.openum.ca/files/sites/82/2021/11/OPC_final.pdf"]}
{"year":"2021","title":"Privacy-Preserving Learning of Human Activity Predictors in Smart Environments","authors":["S Zehtabian, S Khodadadeh, L Bölöni, D Turgut - arXiv preprint arXiv:2101.06564"],"snippet":"… However, many deep learning algorithms work best under big data regimes, where the number of data samples are counted from the tens of thousands (eg the MNIST dataset) to 500 billion (the Common Crawl dataset used to train the GPT-3 model) …","url":["https://arxiv.org/pdf/2101.06564"]}
{"year":"2021","title":"PrivaSeer: A Privacy Policy Search Engine","authors":["S Wilson"],"snippet":"… The URLs in the Common Crawl archive were first filtered based on a selection criteria that took advantage of the fact that most privacy policy URLs either have the word 'privacy' or the words 'data' and 'protection' in them … 6 https://commoncrawl.org …","url":["https://link.springer.com/content/pdf/10.1007/978-3-030-74296-6_22.pdf"]}
{"year":"2021","title":"Probabilistic Text Entry—Case Study 3","authors":["K Vertanen - Intelligent Computing for Interactive System Design …, 2021"],"snippet":"Page 1. 7Probabilistic Text Entry—Case Study 3 Keith Vertanen The reading and writing of text is an activity that has long dominated our interac tions with computers. Early user interfaces often displayed only text and relied on input obtained via a desktop keyboard …","url":["https://dl.acm.org/doi/abs/10.1145/3447404.3447420"]}
{"year":"2021","title":"Probing Multilingual Language Models for Discourse","authors":["M Kurfalı, R Östling - arXiv preprint arXiv:2106.04832, 2021"],"snippet":"… the XLM-tlm. Instead, it is a RoBERTa model (Liu et al., 2019), which is an optimized version of BERT, trained on 2.5 TB of cleaned CommonCrawl data covering 100 languages (Conneau et al., 2020). There are two released …","url":["https://arxiv.org/pdf/2106.04832"]}
{"year":"2021","title":"Proceedings of the First Workshop on Weakly Supervised Learning (WeaSuL)","authors":["MA Hedderich, B Roth, K Kann, B Plank, A Ratner… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. WeaSuL 2021 Proceedings of the First Workshop on Weakly Supervised Learning (WeaSuL) May 7, 2021 Co-located with ICLR (Online) arXiv:2107.03690 v1 [cs.LG] 8 Jul 2021 Page 2. Introduction Welcome to WeaSuL …","url":["https://arxiv.org/pdf/2107.03690"]}
{"year":"2021","title":"Processing large-scale Internet topology data to model Autonomous System Networks","authors":["K Bakhshaliyev - 2020"],"snippet":"… such interrelated data. The number of Facebook users grew from 500 million in 2010 to 1.5 billion in 2014. The Common Crawl web corpora [29] covered for 2012 contains 3.5 billion web pages and 128 billion hyperlinks between …","url":["https://scholarworks.unr.edu/bitstream/handle/11714/7672/Bakhshaliyev_unr_0139D_13330.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Proof Artifact Co-training for Theorem Proving with Language Models","authors":["JM Han, J Rute, Y Wu, EW Ayers, S Polu - arXiv preprint arXiv:2102.06203, 2021"],"snippet":"Page 1. Proof Artifact Co-training for Theorem Proving with Language Models Jesse Michael Han 1 Jason Rute 2 Yuhuai Wu 3 Edward W. Ayers 4 Stanislas Polu 5 Abstract Labeled data for imitation learning of theorem proving …","url":["https://arxiv.org/pdf/2102.06203"]}
{"year":"2021","title":"Propaganda in Press: Challenges of Automatic Detection","authors":["D Pušelj, T Škalec - Text Analysis and Retrieval 2020 Course Project …"],"snippet":"… trained on cased Wikipedia text. RoBERTa has 125M parameters with 12 layers, 768 hidden states, 3072 feed-forward hidden states and 8 heads and is trained on CommonCrawl data in 100 languages. The XLNet is a Large …","url":["https://www.fer.unizg.hr/_download/repository/TAR-2020-ProjectReports.pdf#page=61"]}
{"year":"2021","title":"ProphetNet-X: Large-Scale Pre-training Models for English, Chinese, Multi-lingual, Dialog, and Code Generation","authors":["W Qi, Y Gong, Y Yan, C Xu, B Yao, B Zhou, B Cheng… - arXiv preprint arXiv …, 2021"],"snippet":"… models with downstream task finetuning scripts, including ProphetNet-En pre-trained with 160GB English raw text, ProphetNet-Zh pre-trained with 160GB Chinese raw text, ProphetNet-Multi with 101GB Wiki-100 corpus and …","url":["https://arxiv.org/pdf/2104.08006"]}
{"year":"2021","title":"Proposed Use of a Conversational Agent for Patient Empowerment","authors":["M Alfano, J Kellett, B Lenzitti, M Helfert - 2021"],"snippet":"… Depression 1 https://schema.org/ 2 https://health-lifesci.schema.org/ 3 http:// webdatacommons.org/ 4 http://commoncrawl.org o Dermatitis o Tuberculosis 4.2 Creation of Repository of Schema.org and Health-lifesci Structured Data …","url":["https://www.scitepress.org/Papers/2021/104144/104144.pdf"]}
{"year":"2021","title":"Protecting Anonymous Speech: A Generative Adversarial Network Methodology for Removing Stylistic Indicators in Text","authors":["R Balakrishnan, S Sloan, A Aswani - arXiv preprint arXiv:2110.09495, 2021"],"snippet":"With Internet users constantly leaving a trail of text, whether through blogs, emails, or social media posts, the ability to write and protest anonymously is being eroded because artificial intelligence, when given a sample of previous work, can match text …","url":["https://arxiv.org/pdf/2110.09495"]}
{"year":"2021","title":"Psychographic traits identification based on political ideology: An author analysis study on Spanish politicians' tweets posted in 2020","authors":["JA García-Díaz, R Colomo-Palacios, R Valencia-García - Future Generation …, 2021"],"snippet":"… The pre-trained model of Spanish have been trained from Common Crawl and Wikipedia but adjusted with the PoliCorpus 2020 during training. SE assemble a single vector of dimension 300 for a sequence of the individual word embeddings …","url":["https://www.sciencedirect.com/science/article/pii/S0167739X21004921"]}
{"year":"2021","title":"Putting words into the system's mouth: A targeted attack on neural machine translation using monolingual data poisoning","authors":["J Wang, C Xu, F Guzman, A El-Kishky, Y Tang… - arXiv preprint arXiv …, 2021"],"snippet":"… While this type of attack might appear unrealistic, the nature of the largest collections of monolingual data like Common Crawl (Buck et al., 2014; Wenzek et al., 2020; El-Kishky et al., 2020) (which contains blogs and other …","url":["https://arxiv.org/pdf/2107.05243"]}
{"year":"2021","title":"QMUL-SDS at EXIST: Leveraging Pre-trained Semantics and Lexical Features for Multilingual Sexism Detection in Social Networks","authors":["A Jiang, A Zubiaga - 2021"],"snippet":"… RoBERTa model (XLM-R) [7] has extended the way of pre-training MLM by scaling the amount of data by two orders of magnitude (from Wikipedia to Common Crawl) and training … FastText + TextCNN: we use the Fasttext …","url":["http://www.zubiaga.org/publications/files/jiang2021exist.pdf"]}
{"year":"2021","title":"Quality at a Glance: An Audit of Web-Crawled Multilingual Datasets","authors":["I Caswell, J Kreutzer, L Wang, A Wahab, D van Esch… - arXiv preprint arXiv …, 2021"],"snippet":"… CC: CommonCrawl; TED-6: da, cr, sl, sk, lt, et; TED-45: 45-language subset of (Qi … et al., 2019, 2020) is a set of monolingual corpora extracted from Common Crawl snapshots, specifically from the plain text WET format distributed by Common Crawl which removes …","url":["https://arxiv.org/pdf/2103.12028"]}
{"year":"2021","title":"QuALITY: Question Answering with Long Input Texts, Yes!","authors":["RY Pang, A Parrish, N Joshi, N Nangia, J Phang… - arXiv preprint arXiv …, 2021"],"snippet":"To enable building and testing models on long-document comprehension, we introduce QuALITY, a multiple-choice QA dataset with context passages in English that have an average length of about 5,000 tokens, much longer than typical current …","url":["https://arxiv.org/pdf/2112.08608"]}
{"year":"2021","title":"Quantifying context with and without statistical language models","authors":["CL Jacobs, B Hall"],"snippet":"… Others, such as the Common Crawl (Smith et al., 2013) are a hodge-podge grabbed from across the entire internet, which may include forum posts, blogs, tweets, or computer code, in addition to news articles, fan fiction, and so on …","url":["https://osf.io/mpqut/download"]}
{"year":"2021","title":"Quantum-inspired Hierarchical Attention Mechanism for Question Answering","authors":["P Guo - 2021 IEEE 33rd International Conference on Tools with …, 2021"],"snippet":"… We initialize word embeddings with 300-dimensional Glove word vectors pretrained from the 840B Common Crawl corpus. The LSTM hidden size is set to 100 over WikiQA and TrecQA. AdamOptimizer is used for updating parameters with …","url":["https://ieeexplore.ieee.org/abstract/document/9643371/"]}
{"year":"2021","title":"Query2Prod2Vec Grounded Word Embeddings for eCommerce","authors":["F Bianchi, J Tagliabue, B Yu - arXiv preprint arXiv:2104.02061, 2021"],"snippet":"Page 1. Query2Prod2Vec Grounded Word Embeddings for eCommerce Federico Bianchi Bocconi University Milano, Italy f.bianchi@unibocconi.it Jacopo Tagliabue∗ Coveo Labs New York, USA jtagliabue@coveo.com …","url":["https://arxiv.org/pdf/2104.02061"]}
{"year":"2021","title":"Question Classification for the Travel Domain using Deep Contextualized Word Embedding Models","authors":["C Weerakoon, S Ranathunga - 2021 Moratuwa Engineering Research Conference …, 2021"],"snippet":"… XLNet [5] has pre-trained data over 129 GB collected from BooksCorpus, Wikipedia, CommonCrawl, Giga5, and ClueWeb 2012-B. XLNet at its core uses the permutation technique to predict tokens in random order contrary to masked model of BERT …","url":["https://ieeexplore.ieee.org/abstract/document/9525789/"]}
{"year":"2021","title":"Re-entry Prediction for Online Conversations via Self-Supervised Learning","authors":["L Wang, X Zeng, H Hu, KF Wong, D Jiang - arXiv preprint arXiv:2109.02020, 2021"],"snippet":"… For the parameters in the main model, we first initialize the embedding layer with 200-dimensional Glove embedding (Pennington et al., 2014), whose Twitter version is used for the Twitter dataset and Common Crawl version …","url":["https://arxiv.org/pdf/2109.02020"]}
{"year":"2021","title":"Re-Evaluating GermEval17 Using German Pre-Trained Language Models","authors":["M Aßenmacher, A Corvonato, C Heumann - arXiv preprint arXiv:2102.12330, 2021"],"snippet":"… DBMDZ combine Wikipedia, EU Bookshop (Skadin š et al., 2014), Open Subtitles (Lison and Tiedemann, 2016), CommonCrawl (Ortiz Suárez et al., 2019), ParaCrawl (Espl`a-Gomis et al., 2019) and News Crawl (Haddow …","url":["https://arxiv.org/pdf/2102.12330"]}
{"year":"2021","title":"Read & Improve: A Novel Reading Tutoring System","authors":["R Watson, E Kochmar - EDM 2021: Educational Data Mining, 2021"],"snippet":"… To source news content, R&I monitors both RSS Feeds from news websites and the publicly available Common Crawl News (CC-NEWS) Dataset.3 A fully automated Indexing Pipeline (RIIP, herein) processes article URLs …","url":["https://purehost.bath.ac.uk/ws/files/219929943/EDM_2021.pdf"]}
{"year":"2021","title":"Reading Key Figures from Annual Reports","authors":["S Nordin Hällgren - 2021"],"snippet":"Page 1. Reading Numbers from Annual Reports Using OCR and NLP to Extract Key Figures from Documents Master's thesis in Engineering Mathematics and Computational Science Sara Nordin Hällgren DEPARTMENT OF MATHEMATICAL SCIENCES …","url":["https://odr.chalmers.se/bitstream/20.500.12380/302433/1/Master_thesis_Sara_Nordin_H%C3%A4llgren_210604.pdf"]}
{"year":"2021","title":"Recent advances of low-resource neural machine translation","authors":["R Haque, CH Liu, A Way - Machine Translation, 2021"],"snippet":"… As for the low-resource pairs, monolingual data was provided by the Common Crawl (Buck et al. ) and English monolingual data is blog … 12.Buck C, Heafield K, van Ooyen B (2014) N-gram counts and language models from the common crawl. In …","url":["https://link.springer.com/article/10.1007/s10590-021-09281-1"]}
{"year":"2021","title":"Recent Automatic Post Editing Research","authors":["H Moon, C Park, S Eo, J Seo, H Lim - Journal of Digital Convergence, 2021"],"snippet":"… 진행되었다. 사전학습에는 100개 언어에 대한 Common Crawl[21] 단일 언어 말 뭉치가 이용된다 … 도출하였다. 사전학습에는 25개 언어에 대한 Common Crawl 단 일 언어데이터를 활용하였고, BART[24]에서 제안된 자 기 지도 학습 방법론을 차용하였다 …","url":["https://www.koreascience.or.kr/article/JAKO202123157277846.pdf"]}
{"year":"2021","title":"Recent Research on Phishing Detection Through Machine Learning Algorithm","authors":["A Selamat, O Krejcar - … Conference on Industrial, Engineering and Other …, 2021"],"snippet":"… Phishing URLs, websites or emails are generally collected from PhishTank, Phishing Corpus, OpenPhish, SpamAssassin, etc. [12, 23, 31], whereas legitimate ones are selected from Alexa, Common Crawl, Ham, etc. [16, 24, 28, 31, 35]. Table 4 …","url":["https://link.springer.com/chapter/10.1007/978-3-030-79457-6_42"]}
{"year":"2021","title":"Recognizing and Splitting Conditional Sentences for Automation of Business Processes Management","authors":["NPA Vo, I Manotas, O Popescu, A Cerniauskas… - arXiv preprint arXiv …, 2021"],"snippet":"… length. 4.3 XLM-R-based Model XLM-RoBERTa (XLM-R) is a state of the art multilingual masked language model trained on 2.5 TB of newly created clean CommonCrawl data in 100 languages (Conneau et al., 2019). It obtains …","url":["https://arxiv.org/pdf/2104.00660"]}
{"year":"2021","title":"Reconstructing Implicit Knowledge with Language Models","authors":["M Becker, S Liang, A Frank - Proceedings of Deep Learning Inside Out (DeeLIO) …, 2021"],"snippet":"… GPT-2 is pre-trained on web pages from CommonCrawl, XLNet on CommonCrawl+ ClueWeb (Callan et al., 2009), and BART on the CNN/DM summarization dataset (Hermann et al., 2015). 3.3 Fine-tuning LMs Task-adapted Datasets for LM Fine-tuning …","url":["https://www.aclweb.org/anthology/2021.deelio-1.2.pdf"]}
{"year":"2021","title":"Reducing Human Labor Cost in Deep Learning for Natural Language Processing","authors":["H Jiang - 2021"],"snippet":"… cause it can be trained in a completely unsupervised manner with huge amount of unlabeled data, which are extremely cheap to fetch from internet nowadays. For example, the wellknown “Common Crawl Project”2 is …","url":["https://smartech.gatech.edu/bitstream/handle/1853/64783/JIANG-DISSERTATION-2021.pdf?sequence=1"]}
{"year":"2021","title":"Reference-based Weak Supervision for Answer Sentence Selection using Web Data","authors":["V Krishnamurthy, T Vu, A Moschitti - arXiv preprint arXiv:2104.08943, 2021"],"snippet":"… Work was conducted while the author was an intern at Amazon Alexa AI. more than 100MM Web documents from Common Crawl's data. Given a question-reference pair, the question is used as query to retrieve a set of relevant documents from the index …","url":["https://arxiv.org/pdf/2104.08943"]}
{"year":"2021","title":"Refined Commonsense Knowledge from Large-Scale Web Contents","authors":["TP Nguyen, S Razniewski, J Romero, G Weikum - arXiv preprint arXiv:2112.04596, 2021"],"snippet":"Commonsense knowledge (CSK) about concepts and their properties is useful for AI applications. Prior works like ConceptNet, COMET and others compiled large CSK collections, but are restricted in their expressiveness to subject-predicate-object (SPO) …","url":["https://arxiv.org/pdf/2112.04596"]}
{"year":"2021","title":"Relational Header Discovery using Similarity Search in a Table Corpus","authors":["H Harmouch, T Papenbrock, F Naumann - 2021 IEEE 37th International Conference …, 2021"],"snippet":"… Because web tables contain many distinct values (eg, there are „185M unique values in web tables extracted from the Web Common Crawl [13]), such indexes would be huge and, hence, easily exhaust available memory capacities …","url":["https://ieeexplore.ieee.org/abstract/document/9458838/"]}
{"year":"2021","title":"Reliability Prediction for Health-related Content: A Replicability Study","authors":["M Fernández-Pichel, D Losada, JC Pichel, D Elsweiler - 2021"],"snippet":"… [31] to generate the assessment pool, instead of using a fixed-depth pooling strategy. After the pool was formed, human assessors from 12 http://commoncrawl.org Page 9. Reliability Prediction for Health-related Content 9 Schwarz et al …","url":["https://tec.citius.usc.es/ir/pdf/ECIR_2021_replicability.pdf"]}
{"year":"2021","title":"Representations of Language Varieties Are Reliable Given Corpus Similarity Measures","authors":["J Dunn - arXiv preprint arXiv:2104.01294, 2021"],"snippet":"… The experiments draw on existing geographic corpora from the web (ultimately derived from the Common Crawl data) and from tweets (Dunn and Adams … discussed in Section 4. Data from tweets (TW) is shown on the …","url":["https://arxiv.org/pdf/2104.01294"]}
{"year":"2021","title":"Reproducibility Study: On Heavy-User Bias in A/B Testing","authors":["M Hofmaier, J Schneider, A Wagne"],"snippet":"… they obtained whole Wikipedia to compare it against itself. For the without-analysis they took a 10%-sample of the Common Crawl, which is an open repository for web crawl data. This results in 4.2 million articles consisting of …","url":["https://scholar.archive.org/work/i4ksgrqfh5ekva44jfx6oc433u/access/wayback/https://zenodo.org/record/4459284/files/Reproducibility%20Study%20On%20Heavy-User%20Bias%20in%20AB%20Testing.pdf"]}
{"year":"2021","title":"Research on Recent Quality Estimation","authors":["S Eo, C Park, H Moon, J Seo, H Lim - Journal of the Korea Convergence Society, 2021"],"snippet":"… XLM-R에서는 CommonCrawl[30] 데이터 중 100개 언어에 해당하는 CC100을, mBART에서 25개 언어에 해당하는 CC25를 사전 학습에 활용하였다 … 100개의 언어를 포함하는 2TB 이상의 CommonCrawl 데이터에 …","url":["https://www.koreascience.or.kr/article/JAKO202123162176745.pdf"]}
{"year":"2021","title":"Residual Energy-Based Models for Text","authors":["A Bakhtin, Y Deng, S Gross, M Ott, MA Ranzato… - Journal of Machine …, 2021"],"snippet":"… (2015); Kiros et al. (2015), which consists of fiction books in 16 different genres, totaling about half a billion words. • CCNews: We collect a de-duplicated subset of the English portion of the CommonCrawl news …","url":["https://www.jmlr.org/papers/volume22/20-326/20-326.pdf"]}
{"year":"2021","title":"Retaining Named Entities for Headline Generation","authors":["B Singh, A Marathe, AA Rizvi, AR Joshi - Inventive Computation and Information …"],"snippet":"… These vectors are Page 246. 226 B. Singh et al. trained using word co-occurrence statistics on the Common Crawl corpus. The word embeddings of more than 94% of the words of the dataset's vocabulary are present in the glove. 840B. 300d …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=XhomEAAAQBAJ&oi=fnd&pg=PA221&dq=commoncrawl&ots=EStSRi0ndc&sig=9S5PclK2A3dTQ9RjYbxYB4YNjLI"]}
{"year":"2021","title":"RetroGAN: A Cyclic Post-Specialization System for Improving Out-of-Knowledge and Rare Word Representations","authors":["P Colon-Hernandez, Y Xin, H Lieberman, C Havasi…"],"snippet":"… In our tests we use the English Common Crawl FastText with sub-word information (FT-CC) and Numberbatch 19.08 (NB) to see how performance would be affected by using embeddings that were already retrofitted with a large KB …","url":["https://aclanthology.org/2021.findings-acl.183.pdf"]}
{"year":"2021","title":"Revisiting Additive Compositionality: AND, OR and NOT Operations with Word Embeddings","authors":["M Naito, S Yokoi, G Kim, H Shimodaira - arXiv preprint arXiv:2105.08585, 2021"],"snippet":"Page 1. Revisiting Additive Compositionality: AND, OR and NOT Operations with Word Embeddings Masahiro Naito1,3 Sho Yokoi 2,3 Geewook Kim4 Hidetoshi Shimodaira 1,3 1Kyoto University 2Tohoku University 3RIKEN 4 Naver …","url":["https://arxiv.org/pdf/2105.08585"]}
{"year":"2021","title":"Revisiting Document Representations for Large-Scale Zero-Shot Learning","authors":["J Kil, WL Chao - arXiv preprint arXiv:2104.10355, 2021"],"snippet":"… al., 2009) that has more than 20,000 classes, ex- isting ZSL algorithms (Frome et al., 2013; Norouzi et al., 2013) mostly resort to word vectors of classes names (Mikolov et al., 2013; Pennington et al., 2014) that are automatically …","url":["https://arxiv.org/pdf/2104.10355"]}
{"year":"2021","title":"Revisiting the Primacy of English in Zero-shot Cross-lingual Transfer","authors":["I Turc, K Lee, J Eisenstein, MW Chang, K Toutanova - arXiv preprint arXiv …, 2021"],"snippet":"… with impressive crosslingual transfer capabilities. mT5 Multilingual T5 (mT5) (Xue et al., 2020) is an encoder-decoder model that was jointly trained on 101 languages from Common Crawl. We use its mT5-Base variant, whose …","url":["https://arxiv.org/pdf/2106.16171"]}
{"year":"2021","title":"Revisiting Tri-training of Dependency Parsers","authors":["J Wagner, J Foster - arXiv preprint arXiv:2109.08122, 2021"],"snippet":"… 3.4 Unlabelled Data To match the training data of the word embeddings (Section 3.5), we use the Wikipedia and Common Crawl data of the CoNLL 2017 Shared Task in UD Parsing (Ginter et al., 2017; Zeman et al., 2017) as unlabelled data in tri-training …","url":["https://arxiv.org/pdf/2109.08122"]}
{"year":"2021","title":"Rewarding Open-Domain Conversational Agents Using Memory Networks","authors":["S Majumdar - 2021"],"snippet":"… To collect such pairs of sentences, we use the Common Crawl Monolingual datasets [5]. Hence the final rescaled reward that we use for our experiments is … url: https://commoncrawl.org/. [6] Jacob Devlin et al. “Bert: Pre-training of deep …","url":["https://search.proquest.com/openview/381df303e754ecee58f50ac4b9c84b17/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Rissanen Data Analysis: Examining Dataset Characteristics via Description Length","authors":["E Perez, D Kiela, K Cho - arXiv preprint arXiv:2103.03872, 2021"],"snippet":"Page 1. Rissanen Data Analysis: Examining Dataset Characteristics via Description Length Ethan Perez1 Douwe Kiela2 Kyunghyun Cho1 3 1New York University, 2Facebook AI Research 3CIFAR Fellow in Learning in Machines & Brains perez@nyu.edu Abstract …","url":["https://arxiv.org/pdf/2103.03872"]}
{"year":"2021","title":"Robustness of Machine Translation for Low-Resource Languages","authors":["O Aarnikoivu"],"snippet":"Abstract It is becoming increasingly common for researchers and practitioners to rely on methods within the field of Neural Machine Translation (NMT) that require the use of an extensive amount of auxiliary data. This is especially true for low-resource …","url":["https://project-archive.inf.ed.ac.uk/msc/20215190/msc_proj.pdf"]}
{"year":"2021","title":"Robustness Testing of Language Understanding in Dialog Systems","authors":["J Liu, R Takanobu, J Wen, D Wan, W Nie, H Li, C Li… - arXiv preprint arXiv …, 2020"],"snippet":"… system. We directly use the released models in the repository as DeepSpeech2's configuration, where the speech model is trained on Baidu Internal English Dataset, and the language model is trained on CommonCrawl Data …","url":["https://arxiv.org/pdf/2012.15262"]}
{"year":"2021","title":"RoGPT2: Romanian GPT2 for Text Generation","authors":["MA Niculescu, S Ruseti, M Dascalu - 2021 IEEE 33rd International Conference on …, 2021"],"snippet":"… First, OSCAR [21] has the highest percentage of the total training corpus, is a multi-language corpus, each language being extracted via language classification, filtered, and cleaned from Common Crawl. Second, the latest version of the Wikipedia Dumps for …","url":["https://ieeexplore.ieee.org/abstract/document/9643330/"]}
{"year":"2021","title":"Role of Language Relatedness in Multilingual Fine-tuning of Language Models: A Case Study in Indo-Aryan Languages","authors":["TI Dhamecha, R Murthy V, S Bharadwaj… - arXiv preprint arXiv …, 2021"],"snippet":"… 4.1 Data To train the language models, we obtained text data from various sources including: Wikipedia Dump6, WMT Common Crawl 7, WMT News CommonCrawl8, Urdu Charles University Corpus (Bojar et al., 2014; Jawaid …","url":["https://arxiv.org/pdf/2109.10534"]}
{"year":"2021","title":"ROMANIAN SPEECH RECOGNITION EXPERIMENTS FROM THE ROBIN PROJECT","authors":["AM AVRAM, V PĂIȘ, DAN TUFIȘ - ISSN 1843-911X"],"snippet":"… model could benefit in terms of accuracy. 4 http://corolaws. racai. ro/ corola_sound_search/index. php 5 http://www. racai. ro/p/reterom/ 6 https:// commoncrawl. org/ Page 116. ANDREI-MARIUS AVRAM, VASILE PĂIȘ, DAN TUFIȘ …","url":["https://profs.info.uaic.ro/~consilr/wp-content/uploads/2021/02/volum-ConsILR-v-4-final-revizuit.pdf#page=111"]}
{"year":"2021","title":"RuSentEval: Linguistic Source, Encoder Force!","authors":["V Mikhailov, E Taktasheva, E Sigdel, E Artemova - arXiv preprint arXiv:2103.00573, 2021"],"snippet":"… XLM-R (Conneau et al., 2020a) is trained on 'dy- namic' MLM task, over filtered CommonCrawl data in 100 languages (Wenzek et al., 2020). MiniLM (Wang et al., 2020) is a distilled transformer of BERT architecture, but uses the XLMRoBERTa tokenizer …","url":["https://arxiv.org/pdf/2103.00573"]}
{"year":"2021","title":"S&P500 INDEX DIRECTION PREDICTION USING TEXTUAL TWEETS AND THEIR CORRESPONDING SENTIMENT","authors":["P Mohammadalizadeh, M Jafari - 2021"],"snippet":"… The model structure is a standard sort of encoder-decoder transformer. T5 model structure is shown in Fig. 3. T5 uses common crawl web extracted text and the authors applied a simple heuristic filtering. T5 removes any …","url":["http://journal.research.fanap.com/article_135403_775c4f187c40181e2720e9fc5fc5eaec.pdf"]}
{"year":"2021","title":"S-NLP at SemEval-2021 Task 5: An Analysis of Dual Networks for Sequence Tagging","authors":["VA Nguyen, TM Nguyen, HQ Dao, QH Pham"],"snippet":"… We discard subword part, take only word vector part of a FastText model (pre-trained on Common Crawl dataset) for word representation and utilize an external English Byte Pair Embedding for out-of-vocabulary functionality …","url":["https://aclanthology.org/2021.semeval-1.120.pdf"]}
{"year":"2021","title":"SAFFRON: tranSfer leArning For Food-disease RelatiOn extractioN","authors":["G Cenikj, T Eftimov, BK Seljak - Proceedings of the 20th Workshop on Biomedical …, 2021"],"snippet":"… Wikipedia, which are used for the pretraining of BERT, RoBERTa is trained on data from 3 additional sources: the CommonCrawl News dataset (Nagel, 2016), the OpenWebText corpus (Gokaslan and Cohen, 2019) and Stories …","url":["https://www.aclweb.org/anthology/2021.bionlp-1.4.pdf"]}
{"year":"2021","title":"Samanantar: The Largest Publicly Available Parallel Corpora Collection for 11 Indic Languages","authors":["G Ramesh, S Doddapaneni, A Bheemaraj… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Samanantar: The Largest Publicly Available Parallel Corpora Collection for 11 Indic Languages Gowtham Ramesh1∗ Sumanth Doddapaneni1∗ Aravinth Bheemaraj2,5 Mayank Jobanputra3 …","url":["https://arxiv.org/pdf/2104.05596"]}
{"year":"2021","title":"SAU'S Submission for CCMT 2021 Quality Estimation Task","authors":["Y Li, N Ye, D Cai"],"snippet":"… XLM-R is a multi-language pre-training model proposed by Facebook, which uses 2.5TB CommonCrawl to filter data, and masked language model pre-trained on text in 100 languages, which obtains state-of-the-art performance on cross-lingual …","url":["http://sc.cipsc.org.cn/mt/conference/2021/papers/T21-2012.pdf"]}
{"year":"2021","title":"SAUCE: Truncated Sparse Document Signature Bit-Vectors for Fast Web-Scale Corpus Expansion","authors":["M Wahed, D Gruhl, A Alba, AL Gentile, P Ristoski… - arXiv preprint arXiv …, 2021"],"snippet":"… This experiment facilitates evaluation on a limited seed corpus scenario. The document collection is a snapshot (approximately 200 million documents) of the Common Crawl web archive1, which is regularly updated and consists …","url":["https://arxiv.org/pdf/2108.11948"]}
{"year":"2021","title":"Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning","authors":["N Ram, T Gummadi, R Bhethanabotla, RJ Savery… - Proceedings of the 9th …, 2021"],"snippet":"… The authors opted for an encoder-decoder transformer based design, trained on a variant of the Common Crawl corpus named the Colossal Common Crawl Corpus (or C4 for short). The T5 shows a remarkable ability to tackle many different tasks with …","url":["https://dl.acm.org/doi/abs/10.1145/3472307.3484175"]}
{"year":"2021","title":"Scalable Graph Convolutional Variational Autoencoders","authors":["D Unyi, B Gyires-Tóth - 2021 IEEE 15th International Symposium on Applied …, 2021"],"snippet":"… In Reddit, nodes are Reddit posts, and if the same user commented on two posts, a link is drawn between them; node features are GloVe CommonCrawl word vectors [31] based on the average embedding of the …","url":["https://ieeexplore.ieee.org/abstract/document/9465579/"]}
{"year":"2021","title":"SCALE EFFICIENTLY: INSIGHTS FROM PRE-TRAINING AND FINE-TUNING TRANSFORMERS","authors":["AT Vaswani, D Yogatama, D Metzler, HW Chung… - 2022","FT TRANSFORMERS","Y Tay, M Dehghani, J Rao, W Fedus, S Abnar… - arXiv preprint arXiv …, 2021"],"snippet":"Abstract: There remain many open questions pertaining to the scaling behaviour of Transformer architectures. These scaling decisions and findings can be critical, as training runs often come with an associated computational cost which have both …","url":["https://arxiv.org/pdf/2109.10686","https://openreview.net/pdf?id=f2OYVDyfIB","https://research.google/pubs/pub52066.pdf"]}
{"year":"2021","title":"Scaling Laws for Transfer","authors":["D Hernandez, J Kaplan, T Henighan, S McCandlish - arXiv preprint arXiv:2102.01293, 2021"],"snippet":"… Pre-trained text models were trained on a mix of WebText2 described in [KMH+20], Common Crawl5 [RSR+20], English Wikipedia, and publicly available Internet Books … 5https://commoncrawl.org/the-data/ 6https://www.gharchive.org/ 6 Page 7. 3 Results …","url":["https://arxiv.org/pdf/2102.01293"]}
{"year":"2021","title":"Scarecrow: A Framework for Scrutinizing Machine Text","authors":["Y Dou, M Forbes, R Koncel-Kedziorski, NA Smith… - arXiv preprint arXiv …, 2021"],"snippet":"… GPT-3 DaVinci (Brown et al., 2020) The 175B parameter variant of GPT-3, which is trained on a version of the Common Crawl web scrape with additional filtering and deduplicating. These model choices allow us to study several …","url":["https://arxiv.org/pdf/2107.01294"]}
{"year":"2021","title":"SCOPA: Soft Code-Switching and Pairwise Alignment for Zero-Shot Cross-lingual Transfer","authors":["D Lee, J Lee, G Lee, B Chun, S Hwang - Proceedings of the 30th ACM International …, 2021"],"snippet":"… To enhance such transfer, XLM-R [4] is pre-trained on 100 languages with CommonCrawl Corpora, supervised by translational objectives, to transfer from resource-rich languages (eg, English and Chinese) to resource-poor languages …","url":["https://dl.acm.org/doi/abs/10.1145/3459637.3482176"]}
{"year":"2021","title":"Screening Gender Transfer in Neural Machine Translation","authors":["G Wisniewski, L Zhu, N Bailler, F Yvon - … of the Fourth BlackboxNLP Workshop on …, 2021"],"snippet":"… It includes the Europarl, NewsCommentary and CommonCrawl corpora, and altogether contains 4,813,682 sentences and nearly 141 million French running words. All the corpora were tokenized and segmented into sub-lexical units using …","url":["https://aclanthology.org/2021.blackboxnlp-1.24.pdf"]}
{"year":"2021","title":"Screening hardware and volume factors in distributed machine learning algorithms on spark","authors":["JB Rodrigues, GC Vasconcelos, PRM Maciel - Computing, 2021"],"snippet":"This paper presents an approach to investigate distributed machine learning workloads on Spark. The work analyzes hardware and volume data factors regardin.","url":["https://link.springer.com/article/10.1007/s00607-021-00965-3"]}
{"year":"2021","title":"Selecting Parallel In-domain Sentences for Neural Machine Translation Using Monolingual Texts","authors":["JPR Sharami, D Shterionov, P Spronck - arXiv preprint arXiv:2112.06096, 2021"],"snippet":"Continuously-growing data volumes lead to larger generic models. Specific use-cases are usually left out, since generic models tend to perform poorly in domain-specific cases. Our work addresses this gap with a method for selecting in-domain data from …","url":["https://arxiv.org/pdf/2112.06096"]}
{"year":"2021","title":"Selecting the best data filtering method for NMT training","authors":["F Bane, A Zaretskaya - Proceedings of Machine Translation Summit XVIII …, 2021"],"snippet":"… Rule Based Classifier for Translation Memory Cleaning. Online. Smith, JR, Saint-Amand, H., Plamada, M., Koehn, P., Callison-Burch, C., and Lopez, A. (2013). Dirt cheap web-scale parallel text from the Common Crawl. In …","url":["https://aclanthology.org/2021.mtsummit-up.9.pdf"]}
{"year":"2021","title":"Self-Attention Network for Text Representation Learning","authors":["T Shen - 2021"],"snippet":"Page 1. Self-Attention Network for Text Representation Learning by Tao SHEN Thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy under the supervision of Dist. Prof. Chengqi ZHANG, Prof. Guodong LONG and Dr. Jing JIANG …","url":["https://opus.lib.uts.edu.au/bitstream/10453/149266/2/02Whole.pdf"]}
{"year":"2021","title":"Self-Contextualized Attention for Abusive Language Identification","authors":["H Jarquín-Vásquez, HJ Escalante, M Montes - … of the Ninth International Workshop on …, 2021"],"snippet":"… For word representation we used pre-trained fastText embeddings (Mikolov et al., 2018), trained with subword information on Common Crawl, which have been recognized as useful for this task ac- cording to the study presented in (Corazza et al., 2020) …","url":["https://www.aclweb.org/anthology/2021.socialnlp-1.9.pdf"]}
{"year":"2021","title":"Self-Supervised Representation Learning: Introduction, Advances and Challenges","authors":["L Ericsson, H Gouk, CC Loy, TM Hospedales - arXiv preprint arXiv:2110.09327, 2021"],"snippet":"Self-supervised representation learning methods aim to provide powerful deep feature learning without the requirement of large annotated datasets, thus alleviating the annotation bottleneck that is one of the main barriers to practical deployment of …","url":["https://arxiv.org/pdf/2110.09327"]}
{"year":"2021","title":"Self-Training Pre-Trained Language Models for Zero-and Few-Shot Multi-Dialectal Arabic Sequence Labeling","authors":["M Khalifa, M Abdul-Mageed, K Shaalan - arXiv preprint arXiv:2101.04758, 2021"],"snippet":"… XLM-R also uses Common Crawl for training, which is more likely to have dialectal data than the Arabic Wikipedia (used in mBERT), making it more suited to our work … mBERT, XLM-R was trained on much larger data …","url":["https://arxiv.org/pdf/2101.04758"]}
{"year":"2021","title":"Self-training vs Pre-trained Embeddings for Automatic Essay Scoring","authors":["X Zhou, L Yang, X Fan, G Ren, Y Yang, H Lin - China Conference on Information …, 2021"],"snippet":"… In the training process, we use both the pre-trained embeddings and the randomly initialized self-trained embeddings, where the pre-trained word vector comes from Glove Common Crawl (6B token) and the dimension is …","url":["https://link.springer.com/chapter/10.1007/978-3-030-88189-4_12"]}
{"year":"2021","title":"Self–other moral bias: Evidence from implicit measures and the Word-Embedding Association Test","authors":["MH Li, PW Li, LL Rao - Personality and Individual Differences, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0191886921004864"]}
{"year":"2021","title":"Semantic Detection of Targeted Attacks Using DOC2VEC Embedding","authors":["MS El-Rahmany, EH Mohamed, MH Haggag"],"snippet":"… This model was trained using Common Crawl and Wikipedia. CBOW was used to train these models. In this study, we used dimension 300, character n-grams of length 5, a window of size 5, and 10 negatives. …","url":["https://softcom2020.fesb.unist.hr/pdfs/v17n4_2021-0113_rahmany.pdf"]}
{"year":"2021","title":"Semantic maps and metrics for science Semantic maps and metrics for science using deep transformer encoders","authors":["B Chambers, J Evans - arXiv preprint arXiv:2104.05928, 2021"],"snippet":"Page 1. Semantic maps and metrics for science using deep transformer encoders Brendan Chambers QuillBot Knowledge Lab at the University of Chicago James Evans Department of Sociology and Knowledge Lab, University of Chicago Santa Fe Institute …","url":["https://arxiv.org/pdf/2104.05928"]}
{"year":"2021","title":"Semantic Networks for Engineering Design: State of the Art and Future Directions","authors":["J Han, S Sarica, F Shi, J Luo - Journal of Mechanical Design, 2021"],"snippet":"Page 1. Journal of Mechanical Design 1 Semantic Networks for Engineering Design: State of the Art and Future Directions Ji Han1 School of Engineering University of Liverpool, the United Kingdom e-mail: ji.han@liverpool.ac.uk …","url":["https://asmedigitalcollection.asme.org/mechanicaldesign/article-abstract/doi/10.1115/1.4052148/1115821"]}
{"year":"2021","title":"Semantic Search of Mobile Applications Using Word Embeddings","authors":["J Coelho, A Neto, M Tavares, C Coutinho, R Ribeiro… - 10th Symposium on …, 2021"],"snippet":"… This way, character-level embeddings are considered, instead of word-level representations. In preliminary tests, we compared a CBOW [12] fastText model (FT1), pre-trained on English Common Crawl considering word 5-grams …","url":["https://drops.dagstuhl.de/opus/volltexte/2021/14429/pdf/OASIcs-SLATE-2021-12.pdf"]}
{"year":"2021","title":"Semantic Similarity Based Evaluation for Abstractive News Summarization","authors":["FB Fikri, K Oflazer, B Yanıkoglu - GEM 2021, 2021"],"snippet":"… et al., 2019). mT5 Multilingual T5 (mT5)(Xue et al., 2020) is a variant of T5 model (Raffel et al., 2020) that was pre-trained for 101 languages including Turkish on a new Common Crawl-based dataset. For Turkish summarization …","url":["https://aclanthology.org/2021.gem-1.pdf#page=36"]}
{"year":"2021","title":"Semantic-WER: A Unified Metric for the Evaluation of ASR Transcript for End Usability","authors":["S Roy - arXiv preprint arXiv:2106.02016, 2021"],"snippet":"… A 4-gram language model trained on common crawl dataset is used for decoding 6 7[28] … 4https://github.com/pytorch/fairseq/tree/master/examples/ wav2vec 5https://github.com/facebookresearch/wav2letter/tree/master/recipes …","url":["https://arxiv.org/pdf/2106.02016"]}
{"year":"2021","title":"Semantically Meaningful Sentence Embeddings","authors":["R Deuja - 2021"],"snippet":"Text embedding is an approach used in Natural Language Processing (NLP) to represent words, phrases, sentences, and documents. It is the process of obtaining numeric representations of text to feed into machine learning models as vectors (arrays …","url":["https://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1247&context=computerscidiss"]}
{"year":"2021","title":"SemFace: Pre-training Encoder and Decoder with a Semantic Interface for Neural Machine Translation","authors":["S Ren, L Zhou, S Liu, F Wei, M Zhou, S Ma"],"snippet":"… Table 1: The datasets used in our experiments. Lang: language; Mono: monolingual; Para: parallel; #Sent: number of sentences in the monolingual corpus; NC: NewsCrawl; CC: CommonCrawl; BE: BigEst Estonian corpus; Wiki: Wiki dumps …","url":["https://aclanthology.org/2021.acl-long.348.pdf"]}
{"year":"2021","title":"Semi-automatic Construction of Sight Words Dictionary for Filipino Text Readability","authors":["JM Imperial - Knowledge Management and Acquisition for Intelligent …"],"snippet":"… The Filipino word embedding model has been pre-trained from the articles of Wikipedia and the publicly-available webscraped data of the Common Crawl project using the Continuous Bag-of-Words (CBOW) model [2]. Word …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=oTofEAAAQBAJ&oi=fnd&pg=PA168&dq=commoncrawl&ots=a3xsZgmDzY&sig=HvrePT8d-vC6XeT31rVJoZx_cyI"]}
{"year":"2021","title":"Semi-supervised log-based anomaly detection via probabilistic label estimation","authors":["L Yang, J Chen, Z Wang, W Wang, J Jiang, X Dong… - 2021 IEEE/ACM 43rd …, 2021"],"snippet":"… Then, PLELog adopts the pre-trained word vectors based on Common Crawl Corpus using the FastText algorithm [46] (which can effectively capture the intrinsic relationship among words in natural language), to extract the …","url":["https://ieeexplore.ieee.org/abstract/document/9401970/"]}
{"year":"2021","title":"Sentence encoding for Dialogue Act classification","authors":["N Duran, S Battle, J Smith - Natural Language Engineering, 2021"],"snippet":"In this study, we investigate the process of generating single-sentence representations for the purpose of Dialogue Act (DA) classification, including several aspects of text pre-processing and input representation which are often overlooked …","url":["https://www.cambridge.org/core/journals/natural-language-engineering/article/sentence-encoding-for-dialogue-act-classification/2EF3DC8E57D1019960D18FDE685B1EBA"]}
{"year":"2021","title":"Sentence simplification with ruGPT3","authors":["AA Shatilov, AI Rey"],"snippet":"… One of the latest works [12] uses an unsupervised approach to automatically create training corpora for simplification in multiple languages from raw Common Crawl web data and train simplification systems in any language with a controllable generation mechanism …","url":["http://www.dialog-21.ru/media/5281/shatilovaaplusreyai142.pdf"]}
{"year":"2021","title":"Sentiment Analysis of Code-Mixed Roman Urdu-English Social Media Text using Deep Learning Approaches","authors":["A Younas, R Nasim, S Ali, G Wang, F Qi - 2020 IEEE 23rd International Conference on …"],"snippet":"… Similarly, authors have developed XLM-RoBERTa (XLM-R) - a multilingual language model trained on 2.5 terabyte of filtered CommonCrawl data using BERT-large architecture [9]. It provides strong advantage over mBERT model …","url":["https://ieeexplore.ieee.org/abstract/document/9345887/"]}
{"year":"2021","title":"Sentiment analysis of Korean movie reviews using XLM-R","authors":["NR Shin, TH Kim, DY Yun, SJ Moon, C Hwang - International Journal of Advanced …, 2021"],"snippet":"… Among them, the Facebook AI team unveiled the XLM-R (XLM-RoBERTa), an upgraded XLM model. XLM-R solved the data limitation and the curse of multilinguality by training XLM with 2TB or more refined CC (CommonCrawl), not Wikipedia data …","url":["https://www.koreascience.or.kr/article/JAKO202119559811514.pdf"]}
{"year":"2021","title":"Sentiment Analysis of Latin Poetry: First Experiments on the Odes of Horace","authors":["R Sprugnoli, F Mambrini, M Passarotti, G Moretti - Eighth Italian Conference on …, 2021"],"snippet":"In this paper we present a set of annotated data and the results of a number of unsupervised experiments for the analysis of sentiment in Latin poetry. More specifically, we describe a small gold standard made of eight poems by Horace, in …","url":["http://ceur-ws.org/Vol-3033/paper25.pdf"]}
{"year":"2021","title":"Sentiment Analysis of Sinhala News Comments","authors":["S Ranathunga, IU Liyanage - Transactions on Asian and Low-Resource Language …, 2021"],"snippet":"… 2009], and the Sinhala Common Crawl dataset1 are examples of un-annotated Sinhala corpora. Common crawl is the largest among these, with about 110 million uncleaned tokens … Both SinMin and Common Crawl contain a lot of noise …","url":["https://dl.acm.org/doi/abs/10.1145/3445035"]}
{"year":"2021","title":"Sentiment Analysis of Users' Reviews on COVID-19 Contact Tracing Apps with a Benchmark Dataset","authors":["K Ahmad, F Alam, J Qadir, B Qolomany, I Khan, T Khan… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. 1 Sentiment Analysis of Users' Reviews on COVID-19 Contact Tracing Apps with a Benchmark Dataset Kashif Ahmad∗, Firoj Alam†, Junaid Qadir‡, Basheer Qolomany§, Imran Khan¶, Talhat Khan¶, Muhammad Suleman …","url":["https://arxiv.org/pdf/2103.01196"]}
{"year":"2021","title":"Sentiment Analysis Using XLM-R Transformer and Zero-shot Transfer Learning on Resource-poor Indian Language","authors":["A Kumar, VHC Albuquerque - Transactions on Asian and Low-Resource Language …, 2021"],"snippet":"… It is a scaled cross-lingual sentence encoder trained on 2.5 Tb of data across 100 languages data filtered from CommonCrawl texts. Compared to the original version, the biggest update of XLM-Roberta is a significant …","url":["https://dl.acm.org/doi/abs/10.1145/3461764"]}
{"year":"2021","title":"Sentiment Polarity Classification of Corporate Review Data with a Bidirectional Long-Short Term Memory (biLSTM) Neural Network Architecture","authors":["RE Loke, O Kachaniuk - 2020"],"snippet":"… Therefore, it was decided to proceed with fastText embeddings. Accordingly, pre-trained vectors for Dutch language were added to the biLSTM model as one of the layers. Moreover, these vectors were trained on Common Crawl and Wikipedia using fastText …","url":["https://pure.hva.nl/ws/files/17012226/DATA_2020_75.pdf"]}
{"year":"2021","title":"Sequential Reptile: Inter-Task Gradient Alignment for Multilingual Learning","authors":["S Lee, HB Lee, J Lee, SJ Hwang - arXiv preprint arXiv:2110.02600, 2021"],"snippet":"Page 1. Preprint SEQUENTIAL REPTILE: INTER-TASK GRADIENT ALIGNMENT FOR MULTILINGUAL LEARNING Seanie Lee1∗, Hae Beom Lee1∗, Juho Lee1,2, Sung Ju Hwang1,2 KAIST1, AITRICS2, South Korea …","url":["https://arxiv.org/pdf/2110.02600"]}
{"year":"2021","title":"SG-MLP: Switch Gated Multi-Layer Perceptron Model for Natural Language Understanding","authors":["G Son, S Kim, SJ Joo, W Cho, JE Nah - … of the Korea Information Processing Society …, 2021"],"snippet":"… SG-MLP Small 은 약 16GB 에 해당하는 Book Corpus 와 Wiki 데이터로 학습 되었으며 이외의 모델은 Common Crawl’s Web Crawl Corpus 를 한 차례 정제한 Allen·C4·En[8] 데이터를 학습에 활용하였으며 이는 약 305GB 정도의 분량이다. 보유한 하드웨어 …","url":["https://www.koreascience.or.kr/article/CFKO202133649066979.pdf"]}
{"year":"2021","title":"Shall I Work with Them? A Knowledge Graph-Based Approach for Predicting Future Research Collaborations","authors":["N Kanakaris, N Giarelis, I Siachos, N Karacapilidis - Entropy, 2021"],"snippet":"We consider the prediction of future research collaborations as a link prediction problem applied on a scientific knowledge graph. To the best of our knowledge, this is the first work on the prediction of future research collaborations that …","url":["https://www.mdpi.com/1099-4300/23/6/664/pdf"]}
{"year":"2021","title":"Shallow Discourse Parsing for German","authors":["P Bourgonje"],"snippet":"Page 1. Shallow Discourse Parsing for German Peter Bourgonje Dissertation eingereicht bei der Humanwissenschaftlichen Fakultät der Universität Potsdam 2021 Page 2. Datum der Einreichung: 22-07-2020 Wissenschaftlicher Betreuer und Gutachter: Prof …","url":["https://publishup.uni-potsdam.de/files/50663/bourgonje_diss.pdf"]}
{"year":"2021","title":"Sharp Learning Bounds for Contrastive Unsupervised Representation Learning","authors":["H Bao, Y Nagano, K Nozawa - arXiv preprint arXiv:2110.02501, 2021"],"snippet":"Page 1. Sharp Learning Bounds for Contrastive Unsupervised Representation Learning *Han Bao, Yoshihiro Nagano, Kento Nozawa The University of Tokyo & RIKEN AIP {tsutsumi@ms., nagano@, nozawa@ms.}ku-tokyo.ac.jp October 7, 2021 …","url":["https://arxiv.org/pdf/2110.02501"]}
{"year":"2021","title":"Should we find another model?: Improving Neural Machine Translation Performance with ONE-Piece Tokenization Method without Model Modification","authors":["C Park, S Eo, H Moon, HS Lim - Proceedings of the 2021 Conference of the North …, 2021"],"snippet":"… corpus. In mBART (Liu et al., 2020), the CC25 corpus was composed of a total of 25 languages extracted from CommonCrawl (CC) (Lample and Conneau, 2019; Wenzek et al., 2019) and used for unified vocabulary extraction …","url":["https://www.aclweb.org/anthology/2021.naacl-industry.13.pdf"]}
{"year":"2021","title":"Show, tell and summarise: learning to generate and summarise radiology findings from medical images","authors":["S Singh, S Karimi, K Ho-Shon, L Hamey - Neural Computing and Applications, 2021"],"snippet":"Radiology plays a vital role in health care by viewing the human body for diagnosis, monitoring, and treatment of medical problems. In radiology practice,.","url":["https://link.springer.com/article/10.1007/s00521-021-05943-6"]}
{"year":"2021","title":"SJ_AJ@ DravidianLangTech-EACL2021: Task-Adaptive Pre-Training of Multilingual BERT models for Offensive Language Identification","authors":["SM Jayanthi, A Gupta - arXiv preprint arXiv:2102.01051, 2021"],"snippet":"… mon Crawl data6. The common crawl data consists of textual data both in native scripts as well as in romanized script for some of the languages. Among the three languages in this task, only Tamil has Romanized data in the Common Crawl …","url":["https://arxiv.org/pdf/2102.01051"]}
{"year":"2021","title":"Sketch-Based Creativity Support Tools Using Deep Learning","authors":["F Huang, E Schoop, D Ha, J Nichols, J Canny - Artificial Intelligence for Human …, 2021"],"snippet":"Abstract Sketching is a natural and effective visual communication medium commonly used in creative processes. Recent developments in deep-learning models drastically improved machines’ ability in understanding and generating …","url":["https://link.springer.com/chapter/10.1007/978-3-030-82681-9_12"]}
{"year":"2021","title":"SLAM: A Unified Encoder for Speech and Language Modeling via Speech-Text Joint Pre-Training","authors":["A Bapna, Y Chung, N Wu, A Gulati, Y Jia, JH Clark… - arXiv preprint arXiv …, 2021"],"snippet":"Unsupervised pre-training is now the predominant approach for both text and speech understanding. Self-attention models pre-trained on large amounts of unannotated data have been hugely successful when fine-tuned on downstream …","url":["https://arxiv.org/pdf/2110.10329"]}
{"year":"2021","title":"Slav-NER: the 3rd Cross-lingual Challenge on Recognition, Normalization, Classification, and Linking of Named Entities across Slavic languages","authors":["J Piskorski, B Babych, Z Kancheva, O Kanishcheva… - The 8th Workshop onBalto …, 2021"],"snippet":"Page 1. Slav-NER: the 3rd Cross-lingual Challenge on Recognition, Normalization, Classification, and Linking of Named Entities across Slavic languages Jakub Piskorski,1 Bogdan Babych,2 Zara Kancheva,3 Olga Kanishcheva …","url":["https://tuhat.helsinki.fi/ws/files/161998023/Slav_NER_Shared_Task.pdf"]}
{"year":"2021","title":"SlovakBERT: Slovak Masked Language Model","authors":["M Pikuliak, Š Grivalský, M Konôpka, M Blšták… - arXiv preprint arXiv …, 2021"],"snippet":"… is a suite of RoBERTa-style LMs. The models support 100 languages, including Slovak. Training data are based on CommonCrawl Web-crawled corpus. Slovak part has 23.2 GB (3.5B tokens). The XLM-R models differ in their …","url":["https://arxiv.org/pdf/2109.15254"]}
{"year":"2021","title":"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-Resource Languages","authors":["K Ogueji, Y Zhu, J Lin"],"snippet":"… We select these languages because they are the languages supported by the British Broadcasting Corporation (BBC) News, which was our main source of data.4 We also obtain additional data from the Common Crawl …","url":["https://cs.uwaterloo.ca/~jimmylin/publications/Ogueji_etal_MRL2021.pdf"]}
{"year":"2021","title":"Smart integration of sensors, computer vision and knowledge representation for intelligent monitoring and verbal human-computer interaction","authors":["T Mavropoulos, S Symeonidis, A Tsanousa… - Journal of Intelligent …, 2021"],"snippet":"The details presented in this article revolve around a sophisticated monitoring framework equipped with knowledge representation and computer vision capabi.","url":["https://link.springer.com/article/10.1007/s10844-021-00648-7"]}
{"year":"2021","title":"SOAR: A Synthesis Approach for Data Science API Refactoring","authors":["A Ni, D Ramos, A Yang, I Lynce, V Manquinho… - arXiv preprint arXiv …, 2021"],"snippet":"… is greatly reduced. However, GloVe is trained with Common Crawl [30] which contains raw webpages, which is a mismatch from our domain of textual data (ie, data science and programming). API matching. Given the representation …","url":["https://arxiv.org/pdf/2102.06726"]}
{"year":"2021","title":"Social network analysis using deep learning: applications and schemes","authors":["AM Abbas - Social Network Analysis and Mining, 2021"],"snippet":"Abstract Online social networks (OSNs) are part of daily life of human beings. Millions of users are connected through online social networks. Due to very large number of users and huge amount of data, social network analysis is a challenging …","url":["https://link.springer.com/article/10.1007/s13278-021-00799-z"]}
{"year":"2021","title":"Social Networks Analysis to Retrieve Critical Comments on Online Platforms","authors":["S Bhandari, R Raju - arXiv preprint arXiv:2102.10495, 2021"],"snippet":"… 2019 IEEE International Conference on Healthcare Informatics (ICHI), pp. 1–3, 2019. [13] S. Nagel, “Cc-news.” https://commoncrawl.org/2016/10/news-datasetavailable/, 2016. [14] M. Heidari and JH Jones, “Using bert to extract …","url":["https://arxiv.org/pdf/2102.10495"]}
{"year":"2021","title":"Social Norm Bias: Residual Harms of Fairness-Aware Algorithms","authors":["M Cheng, M De-Arteaga, L Mackey, AT Kalai - arXiv preprint arXiv:2108.11056, 2021"],"snippet":"… We use the dataset1 and task described by De-Arteaga et al. [25]. The dataset, containing 397,340 biographies spanning twenty-eight occupations, is obtained by using regular expressions to filter the Common Crawl for online biographies …","url":["https://arxiv.org/pdf/2108.11056"]}
{"year":"2021","title":"SOLID: A Large-Scale Semi-Supervised Dataset for Offensive Language Identification","authors":["S Rosenthal, P Atanasova, G Karadzhov, M Zampieri… - Proceedings of the Findings …, 2021"],"snippet":"… relations between words. First is an embedding layer initialized with a concatenation of the GloVe 300-dimensional (Pennington et al., 2014) and FastText's Common Crawl 300-dimensional embeddings (Grave et al., 2018). It is …","url":["https://aclanthology.org/2021.findings-acl.80.pdf"]}
{"year":"2021","title":"Spanish Biomedical Crawled Corpus: A Large, Diverse Dataset for Spanish Biomedical Language Models","authors":["CP Carrino, J Armengol-Estapé, OG Bonet… - arXiv preprint arXiv …, 2021"],"snippet":"… The CommonCrawl1 is very a large repository of crawled websites. However, it needs preprocessing to extract the text relevant to the user. The OSCAR corpus [10] was built applying language identification to CommonCrawl …","url":["https://arxiv.org/pdf/2109.07765"]}
{"year":"2021","title":"Sparse Distillation: Speeding Up Text Classification by Using Bigger Models","authors":["Q Ye, M Khabsa, M Lewis, S Wang, X Ren, A Jaech - arXiv preprint arXiv:2110.08536, 2021"],"snippet":"Distilling state-of-the-art transformer models into lightweight student models is an effective way to reduce computation cost at inference time. However, the improved inference speed may be still unsatisfactory for certain time-sensitive applications. In …","url":["https://arxiv.org/pdf/2110.08536"]}
{"year":"2021","title":"Spreadsheet Data Transformation for Ontology Engineering in Petrochemical Equipment Inspection Tasks","authors":["NO Dorodnykh, AY Yurin - … Conference on Intelligent Information Technologies for …, 2021"],"snippet":"… world. The exploration of the Web crawl (“Google Crawl”, “Common Crawl”, “ClueWeb”) discovered hundreds of millions of tables containing relational data [3]. Presumably, billions of valuable facts can be extracted from them …","url":["https://link.springer.com/chapter/10.1007/978-3-030-87178-9_55"]}
{"year":"2021","title":"SQYQP@ Vaxxstance: Stance Detection for the Antivaxxers Movement","authors":["J Calleja, A Méndez - 2021"],"snippet":"… embeddings, Multilingual cased BERT embeddings are employed, which are trained on cased text in the top 104 languages with the largest Wikipedias, as well as XML-RoBERTa embeddings, which are trained on 2.5 TB of newly …","url":["http://ceur-ws.org/Vol-2943/vaxx_paper2.pdf"]}
{"year":"2021","title":"Stacked Embeddings and Multiple Fine-Tuned XLM-RoBERTa Models for Enhanced Hostility Identification","authors":["XLM Fine-Tuned - Combating Online Hostile Posts in Regional …"],"snippet":"… XLM-RoBERTa. XLM-RoBERTa [10] is a large multilingual model trained on the CommonCrawl Dataset. There are two versions: base and large; both have around 250k words in the vocabulary, and the base has 250M parameters, while large has 560M …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=yD8oEAAAQBAJ&oi=fnd&pg=PA224&dq=commoncrawl&ots=ai1kFeEaRU&sig=UkWmp4nmnIiSJ004dO8RQkmOvmk"]}
{"year":"2021","title":"StarryThoughts: Facilitating Diverse Opinion Exploration on Social Issues","authors":["H Kim, H Kim, KJ Jo, J Kim - Proceedings of the ACM on Human-Computer …, 2021"],"snippet":"… After translation, we computed each opinion's embedded vector with the algorithm using pre-trained GloVe word embeddings built upon 42B tokens from Common Crawl [49]. 4.3 Implementation details The front-end of StarryThoughts is implemented with React …","url":["https://dl.acm.org/doi/abs/10.1145/3449140"]}
{"year":"2021","title":"Step-unrolled Denoising Autoencoders for Text Generation","authors":["N Savinov, J Chung, M Binkowski, E Elsen, A Oord - arXiv preprint arXiv:2112.06749, 2021"],"snippet":"… results on unconditional language modeling on the Colossal Cleaned Common Crawl dataset and a dataset of Python code from GitHub. … • We demonstrate good qualitative results for unconditional generation and inpainting on Colossal Clean …","url":["https://arxiv.org/pdf/2112.06749"]}
{"year":"2021","title":"Stepmothers are mean and academics are pretentious: What do pretrained language models learn about you?","authors":["R Choenni, E Shutova, R van Rooij - arXiv preprint arXiv:2109.10052, 2021","WWE Do, FLDT Fine-Tuning"],"snippet":"09/21/21 - In this paper, we investigate what types of stereotypical information are captured by pretrained language models. We present the f...","url":["https://arxiv.org/pdf/2109.10052","https://deepai.org/publication/stepmothers-are-mean-and-academics-are-pretentious-what-do-pretrained-language-models-learn-about-you"]}
{"year":"2021","title":"Stock Volume Prediction Based on Polarity of Tweets, News, and Historical Data Using Deep Learning","authors":["N Jawahar, J Chelladurai, I Sakthivel, B Bajracharya - 2020 2nd International …, 2020"],"snippet":"… token in order to predict the entity of that token. The CNN core model is pre-trained with GloVe vectors on Common Crawl, with 86.43% precision and 86.37% recall for NER. A python script is written that uses Psycopg to extract …","url":["https://dl.acm.org/doi/abs/10.1145/3440054.3440063"]}
{"year":"2021","title":"Storytelling Exhibitions: Identity, Truth and Wonder","authors":["P Hughes - 2021"],"snippet":""}
{"year":"2021","title":"Strategyproof Learning: Building Trustworthy User-Generated Datasets","authors":["S Farhadkhani, R Guerraoui, LN Hoang - arXiv preprint arXiv:2106.02398, 2021"],"snippet":"Page 1. arXiv:2106.02398v1 [cs.LG] 4 Jun 2021 Strategyproof Learning: Building Trustworthy User-Generated Datasets Sadegh Farhadkhani IC School, EPFL Lausanne, Switzerland sadegh.farhadkhani@epfl.ch Rachid Guerraoui …","url":["https://arxiv.org/pdf/2106.02398"]}
{"year":"2021","title":"Streaming cascade-based speech translation leveraged by a direct segmentation model","authors":["J Iranzo-Sánchez, J Jorge, P Baquero-Arnal… - Neural Networks, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0893608021002057"]}
{"year":"2021","title":"Streamlining the Identification of Emerging Tasks in the O* NET System Using Natural Language Processing (NLP): Technical Summary","authors":["JA Dahlke, DJ Putka - 2021"],"snippet":"… In our case, we used embeddings from the Global Vectors for Word Representation (GloVe; Pennington, Socher, & Manning, 2014) algorithm trained on a “common crawl” of the internet consisting of 42 billion tokens (ie, wordlike pieces of text) …","url":["https://www.onetcenter.org/dl_files/EmergingTasksNLP.pdf"]}
{"year":"2021","title":"Structured Text Generation for Spanish Freestyle Battles using Neural Networks","authors":["P Dal Bianco, I Mindlin, L Lanzarini, F Ronchetti… - 2021 XLVII Latin American …, 2021"],"snippet":"… It was trained on a version of the Common Crawl database [6], which consists of around a trillion words. GPT-3 achieves very good results for text generation, via training first on large databases such as the one mentioned and then performing fine-tuning …","url":["https://ieeexplore.ieee.org/abstract/document/9639929/"]}
{"year":"2021","title":"Study of automatic text summarization approaches in different languages","authors":["Y Kumar, K Kaur, S Kaur - Artificial Intelligence Review, 2021"],"snippet":"Nowadays we see huge amount of information is available on both, online and offline sources. For single topic we see hundreds of articles are available, co.","url":["https://link.springer.com/article/10.1007/s10462-021-09964-4"]}
{"year":"2021","title":"Studying the Relationship between Social Position and the Meaning of Work using Siamese BERT Networks and Generalized K-Means Clustering","authors":["S Stewart"],"snippet":"… 1https://www.glassdoor.com 2Specifically, I used the 300-dimensional embeddings with a 1.9 million-token vocabulary that was pretrained on Common Crawl data. 2 Page 3. Review Occupation Flexible working hours …","url":["http://cs230.stanford.edu/projects_fall_2020/reports/57221844.pdf"]}
{"year":"2021","title":"Style Change Detection on Real-World Data using LSTM-powered Attribution Algorithm","authors":["R Deibel, D Löfflad - CLEF, 2021"],"snippet":"… lab [18]. They were trained on Common Crawl and Wikipedia data, using continuous bag of words with position-weights, in dimension 300, with character n-grams of length five, a window of size five and ten negatives. Unlike …","url":["http://ceur-ws.org/Vol-2936/paper-163.pdf"]}
{"year":"2021","title":"SU-NLP at CheckThat! 2021: Check-Worthiness of Turkish Tweets","authors":["B Carik, R Yeniterzi - 2021"],"snippet":"… Loodos_ALBERT5: pre-trained on the same dataset as Loodos_BERT. • mBERT6: multilingual BERT model pre-trained on Wikipedia in 104 languages. • XLM-RoBERTa 7: this cross-lingual model was pretrained on the CommonCrawl corpus in 100 languages …","url":["http://ceur-ws.org/Vol-2936/paper-37.pdf"]}
{"year":"2021","title":"Sub-Subword N-Gram Features for Subword-Level Neural Machine Translation","authors":["A Martinez, K Sudoh, Y Matsumoto - Journal of Natural Language Processing, 2021"],"snippet":"Page 1. IA0248_04martinez (2021-03-05 13:42) General Paper Sub-Subword N-Gram Features for Subword-Level Neural Machine Translation Ander Martinez † , Katsuhito Sudoh † and Yuji Matsumoto †,†† Neural machine …","url":["https://www.jstage.jst.go.jp/article/jnlp/28/1/28_82/_pdf"]}
{"year":"2021","title":"Subword Pooling Makes a Difference","authors":["J Ács, Á Kádár, A Kornai - arXiv preprint arXiv:2102.10864, 2021"],"snippet":"… We pick these two since they are architecturally similar (both have 12 layers and the same hidden size) making our comparison easier. mBERT was trained on Wikipedia while XLM-RoBERTa was trained on CommonCrawl (Wenzek et al., 2020) …","url":["https://arxiv.org/pdf/2102.10864"]}
{"year":"2021","title":"Subwords-Only Alternatives to fastText for Morphologically Rich Languages","authors":["T Ghukasyan, Y Yeshilbashyan, K Avetisyan - Programming and Computer Software, 2021"],"snippet":"… Furthermore, as a result of embedding whole words, fastText models typically weigh a few gigabytes (Facebook's Common Crawl vectors' .bin and .vec files weigh 7.3 GB and 4.5 GB respectively [8]). This especially becomes …","url":["https://link.springer.com/article/10.1134/S0361768821010059"]}
{"year":"2021","title":"Surprise Language Challenge: Developing a Neural Machine Translation System between Pashto and English in Two Months","authors":["A Birch, B Haddow, AVM Barone, J Helcl, J Waldendorf…"],"snippet":"… Finally, we also use the Pashto-English corpus that was submitted by the Bytedance team to the WMT 2020 cleaning shared task (Koehn et al., 2020). For pretraining the German–English model we use exisiting WMT data …","url":["https://aclanthology.org/2021.mtsummit-research.8.pdf"]}
{"year":"2021","title":"Surveillance of COVID-19 Pandemic using Social Media: A Reddit Study in North Carolina","authors":["C Whitfield, Y Liu, M Anwar - arXiv preprint arXiv:2106.04515, 2021"],"snippet":"… The custom NER models trained for this study were based on spaCy's multi-task convolutional neural network (CNN) which was trained using the OntoNotes 6 corpus, and contain GloVe vectors [31] that were trained on Common Crawl …","url":["https://arxiv.org/pdf/2106.04515"]}
{"year":"2021","title":"Survey of Generative Methods for Social Media Analysis","authors":["S Matwin, A Milios, P Prałat, A Soares, F Théberge - arXiv preprint arXiv:2112.07041, 2021"],"snippet":"This survey draws a broad-stroke, panoramic picture of the State of the Art (SoTA) of the research in generative methods for the analysis of social media data. It fills a void, as the existing survey articles are either much narrower in their scope or are dated …","url":["https://arxiv.org/pdf/2112.07041"]}
{"year":"2021","title":"SwissDial: Parallel Multidialectal Corpus of Spoken Swiss German","authors":["P Dogan-Schönberger, J Mäder, T Hofmann - arXiv preprint arXiv:2103.11401, 2021"],"snippet":"… Page 2. 2. Related Work For various high-resource languages, there are parallel text corpora [2, 3, 4] and monolingual text corpora such as [5], Google News, Common Crawl that are used for the tasks of language modeling and machine translation …","url":["https://arxiv.org/pdf/2103.11401"]}
{"year":"2021","title":"Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity","authors":["W Fedus, B Zoph, N Shazeer - arXiv preprint arXiv:2101.03961, 2021"],"snippet":"Page 1. SWITCH TRANSFORMERS: SCALING TO TRILLION PARAMETER MODELS WITH SIMPLE AND EFFICIENT SPARSITY William Fedus∗ Google Brain liamfedus@google.com Barret Zoph∗ Google Brain barretzoph@google.com …","url":["https://arxiv.org/pdf/2101.03961"]}
{"year":"2021","title":"SYMBOLIC AND NEURAL APPROACHES TO NATURAL LANGUAGE INFERENCE","authors":["H Hu - 2021"],"snippet":"Page 1. SYMBOLIC AND NEURAL APPROACHES TO NATURAL LANGUAGE INFERENCE Hai Hu Submitted to the faculty of the University Graduate School in partial fulfillment of the requirements for the degree …","url":["https://scholarworks.iu.edu/dspace/bitstream/handle/2022/26642/dissertation_final_hai_hu.pdf?sequence=1"]}
{"year":"2021","title":"Symbolic Knowledge Distillation: from General Language Models to Commonsense Models","authors":["P West, C Bhagavatula, J Hessel, JD Hwang, L Jiang… - arXiv preprint arXiv …, 2021"],"snippet":"The common practice for training commonsense models has gone from-human-to-corpus-to-machine: humans author commonsense knowledge graphs in order to train commonsense models. In this work, we investigate an alternative, from-machine-to-corpus-to-machine …","url":["https://arxiv.org/pdf/2110.07178"]}
{"year":"2021","title":"Syntax-Enhanced Pre-trained Model","authors":["Z Xu, D Guo, D Tang, Q Su, L Shou, M Gong, W Zhong… - arXiv preprint arXiv …, 2020"],"snippet":"… pre-training and fine-tuning stages. We construct a pre-training dataset by applying an offthe-shell dependency parser (Qi et al., 2020) to one billion sentences from common crawl news. With these data, we introduce a syntax …","url":["https://arxiv.org/pdf/2012.14116"]}
{"year":"2021","title":"Systemic racism: individuals and interactions, institutions and society","authors":["MR Banaji, ST Fiske, DS Massey - Cognitive Research: Principles and Implications, 2021","ST Fiske, DS Massey - Cognitive Research, 2021"],"snippet":"Systemic racism is a scientifically tractable phenomenon, urgent for cognitive scientists to address. This tutorial reviews the built-in systems that undermine life opportunities and outcomes by racial category, with a focus on challenges to Black …","url":["https://cognitiveresearchjournal.springeropen.com/articles/10.1186/s41235-021-00349-3","https://search.proquest.com/openview/05ee76c9a6696a3ec2aa4bcd2f966d1d/1?pq-origsite=gscholar&cbl=4402895"]}
{"year":"2021","title":"Tab2KG: Semantic Table Interpretation with Lightweight Semantic Profiles","authors":["S Gottschalk, E Demidova"],"snippet":"… Subsequent steps such as property mapping are based on the results of this lookup step. However, as shown by Ritze et al. [14], only about 3% of the tables contained in the 3.5 billion HTML pages of the Common …","url":["http://semantic-web-journal.net/system/files/swj2731.pdf"]}
{"year":"2021","title":"TABBIE: Pretrained Representations of Tabular Data","authors":["H Iida, D Thai, V Manjunatha, M Iyyer - arXiv preprint arXiv:2105.02584, 2021"],"snippet":"… TaBERT's pretraining data was not publicly released at the time of our work, but their dataset consists of 26.6M ta- bles from Wikipedia and the Common Crawl. We thus form a pretraining dataset of equivalent size by combining …","url":["https://arxiv.org/pdf/2105.02584"]}
{"year":"2021","title":"Tabular Data Concept Type Detection Using Star-Transformers","authors":["Y Zhou, S Singh, C Christodoulopoulos - 2021"],"snippet":"… We directly fine-tune TURL's concept type detection model on our dataset using its re- leased code. 4.2 Training Details We use 300 dimensional Glove [12] word embeddings pre-trained on 42 billion uncased tokens from Common …","url":["https://assets.amazon.science/80/73/db622f0344878705986e0a99dc1f/tabular-data-concept-type-detection-using-star-transformers.pdf"]}
{"year":"2021","title":"Tackling Cyber-Aggression: Identification and Fine-Grained Categorization of Aggressive Texts on Social Media using Weighted Ensemble of Transformers","authors":["O Sharif, MM Hoque - Neurocomputing, 2021"],"snippet":"… This model is trained with monolingual Bengali CommonCrawl corpus and utilises the BERT base model’s architecture. XLM-R is a cross-lingual model which outdoes m-BERT in various benchmarks. This model is built over the of 100 …","url":["https://www.sciencedirect.com/science/article/pii/S0925231221018567"]}
{"year":"2021","title":"Tackling Italian University Assessment Tests with Transformer-Based Language Models","authors":["D Puccinelli, S Demartini, PL Ferrari - 2021"],"snippet":"Cloze tests are a great tool to asses reading proficiency as well as analytical thinking, and are therefore employed in admission and assessment tests at various levels of the education system in multiple countries. In Italy, cloze tests are administered to …","url":["http://ceur-ws.org/Vol-3033/paper20.pdf"]}
{"year":"2021","title":"Targeted Feedback Generation for Constructed-Response Questions","authors":["J Zhao, K Larson, W Xu, N Gattani, C Thille - 2020"],"snippet":"… TextRank, BERT basic, and BERT fine-tuned models achieve better FPR scores. For non-transformer-based approaches, we ap- ply multitask CNN (trained on OntoNotes) and GloVe (1.2m 300d vectors trained on Common Crawl) as our embedding methods …","url":["https://assets.amazon.science/bb/f6/ea2108414befb188d20e2b75ac87/targeted-feedback-generation-for-constructed-response-questions.pdf"]}
{"year":"2021","title":"Task Aware Multi-Task Learning for Speech to Text Tasks","authors":["S Indurthi, MA Zaidi, NK Lakumarapu, B Lee, H Han… - ICASSP 2021-2021 IEEE …, 2021"],"snippet":"… We use the following datasets for ST English-German, ASR English, MT English-German tasks: MT En-De: We use the Open Subtitles [9] and WMT 19 corpora. WMT 19 consists of Common Crawl, Europarl v9, and News …","url":["https://ieeexplore.ieee.org/abstract/document/9414703/"]}
{"year":"2021","title":"Taxonomy Enrichment with Text and Graph Vector Representations","authors":["I Nikishina, M Tikhomirov, V Logacheva, Y Nazarov…"],"snippet":"Page 1. Semantic Web 0 (0) 1 1 IOS Press 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 …","url":["http://www.semantic-web-journal.net/system/files/swj2729.pdf"]}
{"year":"2021","title":"TBCOV: Two Billion Multilingual COVID-19 Tweets with Sentiment, Entity, Geo, and Gender Labels","authors":["M Imran, U Qazi, F Ofli - arXiv preprint arXiv:2110.03664, 2021"],"snippet":"… Although the original XLM-R model is trained on one hundred languages using more than two terabytes of filtered CommonCrawl data4, its Twitter variant XLM-T achieves better performance on a large multilingual benchmark for sentiment analysis17 …","url":["https://arxiv.org/pdf/2110.03664"]}
{"year":"2021","title":"Technology-Skill Complementarity and Labor Displacement: Evidence from Linking Two Centuries of Patents with Occupations","authors":["L Kogan, D Papanikolaou, LDW Schmidt, B Seegmiller"],"snippet":"Abstract We construct new technology indicators using textual analysis of patent documents and occupation task descriptions that span of two centuries (1850–2010). At the industry level, improvements in technology are associated with higher labor …","url":["https://www.bryanseegmiller.com/files/Draft_v2021-1017.pdf"]}
{"year":"2021","title":"Temporal Effects on Pre-trained Models for Language Processing Tasks","authors":["O Agarwal, A Nenkova - arXiv preprint arXiv:2111.12790, 2021"],"snippet":"Keeping the performance of language technologies optimal as time passes is of great practical interest. Here we survey prior work concerned with the effect of time on system performance, establishing more nuanced terminology for discussing the …","url":["https://arxiv.org/pdf/2111.12790"]}
{"year":"2021","title":"Term-community-based topic detection with variable resolution","authors":["A Hamm, S Odrowski - arXiv preprint arXiv:2103.13550, 2021"],"snippet":"… The pre-trained fastText models are shallow neural nets with output vectors of dimension 300 that were trained on huge text collections from Wikipedia and Common Crawl, using the so called CBOW task of predicting a word by its surrounding words …","url":["https://arxiv.org/pdf/2103.13550"]}
{"year":"2021","title":"Term-Community-Based Topic Detection with Variable Resolution. Information 2021, 12, 221","authors":["A Hamm, S Odrowski - 2021"],"snippet":"Page 1. information Article Term-Community-Based Topic Detection with Variable Resolution Andreas Hamm * and Simon Odrowski Citation: Hamm, A.; Odrowski, S. Term-Community-Based Topic Detection with …","url":["https://128.84.4.13/pdf/2103.13550.pdf"]}
{"year":"2021","title":"Text Analysis for Psychology: Methods, Principles, and Practices","authors":["B Kennedy, A Ashokkumar, RL Boyd, M Dehghani - 2021"],"snippet":"Page 1. Running Head: TEXT ANALYSIS METHODS FOR PSYCHOLOGY Text Analysis for Psychology: Methods, Principles, and Practices Brendan Kennedy​1​, Ashwini Ashokkumar​2​, Ryan Boyd​3–5​, Morteza …","url":["https://psyarxiv.com/h2b8t/download?format=pdf"]}
{"year":"2021","title":"Text Classification for Predicting Multi-level Product Categories","authors":["H Jahanshahi, O Ozyegen, M Cevik, B Bulut, D Yigit… - arXiv preprint arXiv …, 2021"],"snippet":"… next-sentence pretraining objectives, and training with larger batch sizes. For our experiments, we use the multi-lingual version of this architecture pretrained on 2.5TB of CommonCrawl data in 100 languages using a …","url":["https://arxiv.org/pdf/2109.01084"]}
{"year":"2021","title":"Text Normalization for Low-Resource Languages of Africa","authors":["A Zupon, E Crew, S Ritchie - arXiv preprint arXiv:2103.15845, 2021"],"snippet":"… The size of available corpora also varies greatly from 10K sentences (Swahili Newscrawl) to 1M sentences (Afrikaans Web and Mixed). The OSCAR9 corpus is a multilingual corpus of data obtained by language classification and filtering of the Common Crawl corpus …","url":["https://arxiv.org/pdf/2103.15845"]}
{"year":"2021","title":"Text Similarity Techniques for Matching Employee Objectives","authors":["A Elnaggar, M Ghanem, F Matthes, AD Mckinnon…"],"snippet":"… https://huggingface.co/albert-xxlarge-v2 11. https://huggingface.co/distilbertbase-uncased 12. https://huggingface.co/roberta-base 13. https:// commoncrawl.org/2016/10/news-dataset-available/ 14. https://github.com …","url":["https://wwwmatthes.in.tum.de/file/wpu1f6pvjvt4/Sebis-Public-Website/-/Guided-Research-Mohab-Ghanem/Text_Similarity_Techniques_For_Matching_Employee_Objective.pdf"]}
{"year":"2021","title":"Text2Gestures: A Transformer-Based Network for Generating Emotive Body Gestures for Virtual Agents","authors":["U Bhattacharya, N Rewkowski, A Banerjee, P Guhan… - arXiv preprint arXiv …, 2021"],"snippet":"… W = [w1 ... ws ... wTsen ], with Tsen being the maximum sentence length, using word embeddings ws ∈ R300. We obtain the word em- beddings using the GloVe model pre-trained on the Common Crawl corpus [52]. We opt …","url":["https://arxiv.org/pdf/2101.11101"]}
{"year":"2021","title":"Text2Gestures: A Transformer-Based Network for Generating Emotive Body Gestures for Virtual Agents** This work has been supported in part by ARO Grants …","authors":["U Bhattacharya, N Rewkowski, A Banerjee, P Guhan… - 2021 IEEE Virtual Reality …, 2021"],"snippet":"… W = [w1 ... ws ... wTsen], with Tsen being the maximum sentence length, using word embeddings ws ∈ R300. We obtain the word embeddings using the GloVe model pre-trained on the Common Crawl corpus [52]. We opt …","url":["https://ieeexplore.ieee.org/abstract/document/9417647/"]}
{"year":"2021","title":"TextSETTR: Few-Shot Text Style Extraction and Tunable Targeted Restyling","authors":["P Rileya, N Constantb, M Guob, G Kumarc, D Uthusb…"],"snippet":"… For this purpose, we train a single general-purpose TextSETTR model, with the same configuration as our model from Section 3, except fine-tuned for 200k steps on English Common Crawl data (the same “C4” data that T5 pretrained on) instead of Amazon reviews …","url":["https://aclanthology.org/2021.acl-long.293.pdf"]}
{"year":"2021","title":"Textual Analysis of News for Stock Market Prediction","authors":["AV Bogdanov, M Bogan, A Stankus - … Conference on Computational Science and Its …, 2021"],"snippet":"… So, for Word2Vec there are pre-trained word-vectors on Google news with a dimension of 300. There is a dictionary for GloVe trained on various textual data, including on news with Common Crawl with a dimension of 300 …","url":["https://link.springer.com/chapter/10.1007/978-3-030-87010-2_22"]}
{"year":"2021","title":"TGIF: Tree-Graph Integrated-Format Parser for Enhanced UD with Two-Stage Generic-to Individual-Language Finetuning","authors":["T Shi, L Lee - arXiv preprint arXiv:2107.06907, 2021"],"snippet":"… Representation The first step is to extract contextual representations. For this purpose, we use the pre-trained XLM-R model (Conneau et al., 2020), which is trained on multilingual CommonCrawl data and supports all 17 languages in the shared task …","url":["https://arxiv.org/pdf/2107.06907"]}
{"year":"2021","title":"The 2020 Bilingual, Bi-Directional WebNLG+ Shared Task Overview and Evaluation Results (WebNLG","authors":["T Ferreira, C Gardent, N Ilinykh, C van der Lee, S Mille…"],"snippet":"… 3.3 Mono-task, Bilingual Approaches cuni-ufal. The mBART model (Liu et al., 2020) is pre-trained for multilingual denoising on the large-scale multilingual CC25 corpus extracted from Common Crawl, which contains …","url":["https://pure.uvt.nl/ws/portalfiles/portal/48450135/2020.webnlg_1.7.pdf"]}
{"year":"2021","title":"The Advantages of Human Evaluation of Sociomedical Question Answering Systems","authors":["V Firsanova - International Journal of Open Information Technologies, 2021"],"snippet":"… XLM-RoBERTa is a model for one hundred languages trained on CommonCrawl data. According to the developers, this model improved the performance on low-resource languages of other crossand multilingual models developed earlier. This model is …","url":["http://injoit.ru/index.php/j1/article/download/1227/1166"]}
{"year":"2021","title":"The Croatian psycholinguistic database: Estimates for 6000 nouns, verbs, adjectives and adverbs","authors":["A Peti-Stantić, M Anđel, V Gnjidić, G Keresteš… - Behavior Research Methods, 2021"],"snippet":"Psycholinguistic databases containing ratings of concreteness, imageability, age of acquisition, and subjective frequency are used in psycholinguistic and.","url":["https://link.springer.com/article/10.3758/s13428-020-01533-x"]}
{"year":"2021","title":"The Danish Gigaword Corpus","authors":["L Derczynski, ITU Copenhagen, MR Ciosici, R Baglini…"],"snippet":"… Massive, monolithic, automatically collected datasets of web content, such as Common Crawl, support the training of large language models but suffer from quality issues (Radford et al., 2019) and bias (Ferrer et al., 2021) …","url":["http://derczynski.com/papers/dagw.pdf"]}
{"year":"2021","title":"The Data Set Knowledge Graph: Creating a Linked Open Data Source for Data Sets","authors":["M Färber, D Lamprecht - 2021"],"snippet":"ABSTRACT Several scholarly knowledge graphs have been proposed to model and analyze the academic landscape. However, although the number of data sets has increased remarkably in recent years, these knowledge graphs do not primarily …","url":["https://www.aifb.kit.edu/images/6/66/DSKG_QSS2021_v0.pdf"]}
{"year":"2021","title":"The Demand for Executive Skills","authors":["J Fuller, S Hansen, T Ramdas, R Sadun - 2021"],"snippet":"… tions. 'Off-the-shelf' models—typically estimated on large corpora that are representative of written language such as Wikipedia and Common Crawl for English—do exist, but se- mantic relatedness in generic English may …","url":["https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3886751"]}
{"year":"2021","title":"The Effect of Domain and Diacritics in Yorubá–English Neural Machine Translation","authors":["DI Adelani, D Ruiter, JO Alabi, D Adebonojo, A Ayeni… - Proceedings of Machine …, 2021"],"snippet":"… Fine-tuning mT5 We examine a transfer learning approach by fine-tuning a massively multilingual model mT5 (Xue et al., 2021). mT5 had been pre-trained on 6.3T tokens originating from Common Crawl in 101 languages (including Yor`ubá) …","url":["https://www.cs.upc.edu/~cristinae/CV/docs/2021.mtsummit.adelani.pdf"]}
{"year":"2021","title":"The Effectiveness of Intermediate-Task Training for Code-Switched Natural Language Understanding","authors":["A Prasad, MA Rehan, S Pathak, P Jyothi - arXiv preprint arXiv:2107.09931, 2021"],"snippet":"… XLM-R uses a similar training objective as mBERT but is trained on orders of magnitude more data from the CommonCrawl corpus spanning 100 languages and yields competitive results on low-resource languages(Conneau et al., 2020) …","url":["https://arxiv.org/pdf/2107.09931"]}
{"year":"2021","title":"The FinSim-2 2021 Shared Task: Learning Semantic Similarities for the Financial Domain","authors":["Y Mansar, J Kang, IE Maarouf - Companion Proceedings of the Web Conference 2021, 2021"],"snippet":"… where english words of similar meaning are grouped into sets called synsets, Wikidata is a knowledge base maintained by the Wikimedia Foundation and contains entities like \"MSCI World\" and finally WebIsALOD is a hypernymy …","url":["https://dl.acm.org/doi/abs/10.1145/3442442.3451381"]}
{"year":"2021","title":"The Impact of Main Content Extraction on Near-Duplicate Detection","authors":["M Fröbe, M Hagen, J Bevendorff, M Völske, B Stein… - arXiv preprint arXiv …, 2021"],"snippet":"… Widely available web crawls—most notably the ClueWebs† and the Common Crawl‡—contain the (near-)duplicate documents that the crawler encountered during the crawling process. While the inclusion of near-duplicates enables many …","url":["https://arxiv.org/pdf/2111.10864"]}
{"year":"2021","title":"The Impact of Positional Encodings on Multilingual Compression","authors":["V Ravishankar, A Søgaard - arXiv preprint arXiv:2109.05388, 2021"],"snippet":"… The second corpus uses the English Wikipedia as the training split, and Common Crawl as validation. We present corpus statistics in Table 3. For each corpus, we learn and apply a BPE vocabulary of size 2048. Train Validation Bible 30602 9080 Wikipedia 50000 20000 …","url":["https://arxiv.org/pdf/2109.05388"]}
{"year":"2021","title":"The information retrieval anthology 2021: inaugural status report and challenges ahead","authors":["M Potthast, B Stein, M Hagen - ACM SIGIR Forum, 2021"],"snippet":"… Originally developed as a reproducible baseline search engine for the ClueWeb09, the ClueWeb12, and the Common Crawl, ChatNoir's distributed Elasticsearch backend also enables the inclusion of other kinds of indexes …","url":["https://dl.acm.org/doi/abs/10.1145/3476415.3476417"]}
{"year":"2021","title":"The Information Retrieval Anthology","authors":["M Potthast, S Günther, J Bevendorff, JP Bittner… - 2021"],"snippet":"… To allow users to easily search the IR Anthology, we provide a dedicated search index accessible via our search engine ChatNoir [6], a proven web search engine indexing around 5 billion web pages from the ClueWeb crawls and Common Crawl versions …","url":["https://webis.de/downloads/publications/papers/potthast_2021j.pdf"]}
{"year":"2021","title":"The IWSLT 2021 BUT Speech Translation Systems","authors":["HK Vydana, M Karafi'at, L Burget - arXiv preprint arXiv:2107.06155, 2021"],"snippet":"… Corpora #Sentences Audio Source text Target Text MT -Train-set ParaCrawl v3 31M - OpenSubtitles 2018 12M - Rapid 2019 1.5M - Europarl v9 1.81M - News Commentary 365K - Common Crawl 2.4M - Wikititles …","url":["https://arxiv.org/pdf/2107.06155"]}
{"year":"2021","title":"The Language of Situational Empathy","authors":["KE ZHOU, LM AIELLO, S SCEPANOVIC, D QUERCIA… - 2021"],"snippet":"… Specifically, we averaged the 300-dimensional embedding vectors for each lexicon word using GloVe embeddings [55] trained on the Common Crawl corpus8. The above two approaches used a machine learning algorithm on top of our lexicon for the final prediction …","url":["https://social-dynamics.net/docs/empathy.pdf"]}
{"year":"2021","title":"The NiuTrans End-to-End Speech Translation System\\\\for IWSLT 2021 Offline Task","authors":["C Xu, X Liu, X Liu, L Wang, C Huang, T Xiao, J Zhu - arXiv preprint arXiv:2107.02444, 2021"],"snippet":"… Task Corpora Size Time ASR LibriSpeech 281241 960h Common Voice 1000000 1387h Total 1281241 2347h MT CommonCrawl 2014304 - Europarl 1802849 - ParaCrawl 31528317 - Wiki 5714363 - OpenSubtitles 14449099 - Total 55508932 - ST …","url":["https://arxiv.org/pdf/2107.02444"]}
{"year":"2021","title":"The NiuTrans Machine Translation Systems for WMT21","authors":["S Zhou, T Zhou, B Wei, Y Luo, Y Mu, Z Zhou, C Wang… - arXiv preprint arXiv …, 2021"],"snippet":"… For the ParaCrawl v7.1, we only selected 8.5 million data according to the score of sentences provided by the dataset. We chose all of News Crawl and News Commentary and 12 million data sampled from Common Crawl for the Japanese monolingual data …","url":["https://arxiv.org/pdf/2109.10485"]}
{"year":"2021","title":"The Pile: An 800GB Dataset of Diverse Text for Language Modeling","authors":["L Gao, S Biderman, S Black, L Golding, T Hoppe… - arXiv preprint arXiv …, 2020"],"snippet":"… The growing need for data in language modeling has caused most existing large-scale language models to turn to the Common Crawl for most or all of their data (Brown et al., 2020; Raffel et al., 2019). While training on the …","url":["https://arxiv.org/pdf/2101.00027"]}
{"year":"2021","title":"The Potential of Automated Text Analytics in Social Knowledge Building","authors":["R Németh, J Koltai - Pathways Between Social Science and Computational …, 2021"],"snippet":"… Finally, big data are often samples themselves, resulting from an unknown or hard-to-formalize sampling procedure. See, eg, Common Crawl (http://commoncrawl.org), an open repository of textual web page data …","url":["https://link.springer.com/chapter/10.1007/978-3-030-54936-7_3"]}
{"year":"2021","title":"The Power of Anchor Text in the Neural Retrieval Era","authors":["M Fröbe, S Günther, M Probst, M Potthast, M Hagen"],"snippet":"… and Eiron and McCurley on MS MARCO, we extract anchor texts from Common Crawl snapshots. We randomly select one snapshot from each year from 2016 to 2021 (each containing 1.7–3.4 billion documents) and extract the anchor texts of …","url":["https://webis.de/downloads/publications/papers/froebe_2022a.pdf"]}
{"year":"2021","title":"The practical ethics of bias reduction in machine translation: why domain adaptation is better than data debiasing","authors":["M Tomalin, B Byrne, S Concannon, D Saunders… - Ethics and Information …, 2021"],"snippet":"This article probes the practical ethical implications of AI system design by reconsidering the important topic of bias in the datasets used to train auton.","url":["https://link.springer.com/article/10.1007/s10676-021-09583-1"]}
{"year":"2021","title":"The Rediscovery Hypothesis: Language Models Need to Meet Linguistics","authors":["T Maxat - 2022","V Nikoulina, M Tezekbayev, N Kozhakhmet… - arXiv preprint arXiv …, 2021"],"snippet":"… English-to-German translation. The authors used the CommonCrawl-840B GloVe model (Pennington et al., 2014) for English word vectors, which was completely fixed during pretraining, and we follow their setup. This entails that …","url":["https://arxiv.org/pdf/2103.01819","https://nur.nu.edu.kz/bitstream/handle/123456789/6141/Tezekbayev,%20Maxat.pdf?sequence=1"]}
{"year":"2021","title":"The SPECTRANS System Description for the WMT21 Terminology Task","authors":["NBDCB Faye, ZYKHM Mojca, PJBYG Wisniewski…"],"snippet":"… The latter did better for sacreBLEU and Window Overlap Accuracy (n=3) but probably having seen the terminological resources in the training data gave an edge for Exact-Match Accuracy to our model trained with Common Crawl and the terminological …","url":["http://www.statmt.org/wmt21/pdf/2021.wmt-1.80.pdf"]}
{"year":"2021","title":"The Trade-offs of Domain Adaptation for Neural Language Models","authors":["D Iter, D Grangier - arXiv preprint arXiv:2109.10274, 2021"],"snippet":"Page 1. The Trade-offs of Domain Adaptation for Neural Language Models David Grangier Google, Mountain View, CA grangier@google.com Dan Iter∗ Stanford, Palo Alto, CA daniter@stanford.edu Abstract This work connects …","url":["https://arxiv.org/pdf/2109.10274"]}
{"year":"2021","title":"The University of Edinburgh's Bengali-Hindi Submissions to the WMT21 News Translation Task","authors":["P Pal, AF Aji, P Chen, S Sen"],"snippet":"Abstract We describe the University of Edinburgh’s Bengali↔ Hindi constrained systems submitted to the WMT21 News Translation task. We submitted ensembles of Transformer models built with large-scale back-translation and finetuned on subsets …","url":["https://www.statmt.org/wmt21/pdf/2021.wmt-1.16.pdf"]}
{"year":"2021","title":"The University of Edinburgh's English-German and English-Hausa Submissions to the WMT21 News Translation Task","authors":["PCJHU Germann, LBN Bogoychev…"],"snippet":"… Page 2. • Crawled parallel: ParaCrawl, WikiMatrix, CommonCrawl, and WikiTitles … For back-translation, we use 9.5 million monolingual Hausa sentences from Common Crawl, Extended Common Crawl, and News Crawl provided by the task organizers …","url":["https://pinzhenchen.github.io/paper/2021-wmt-deen.pdf"]}
{"year":"2021","title":"The USTC-NELSLIP Systems for Simultaneous Speech Translation Task at IWSLT 2021","authors":["D Liu, M Du, X Li, Y Hu, L Dai - arXiv preprint arXiv:2107.00279, 2021"],"snippet":"… sentences from it. Corpus Sentences EN→DE MuST-C(v2) 229.7k Europarl 1828.5k Rapid-2019 1531.3k WIT3-TED 209.5k Commoncrawl 2399.1k WikiMatrix 6227.2k Wikititles 1382.6k Paracrawl 82638.2k EN→JA WIT3-TED 225.0 …","url":["https://arxiv.org/pdf/2107.00279"]}
{"year":"2021","title":"The USYD-JD Speech Translation System for IWSLT 2021","authors":["L Ding, D Wu, D Tao - arXiv preprint arXiv:2107.11572, 2021"],"snippet":"… Corpus #Sent. commoncrawl English 4,366,344 commoncrawl Swahili 38,928 + upsampling (14×) 544,992 … Mono. Corpus for Tagged BT #Sent. Totally collected corpus commoncrawl English 30,513,498 Cleaned corpus …","url":["https://arxiv.org/pdf/2107.11572"]}
{"year":"2021","title":"The Utility and Interplay of Gazetteers and Entity Segmentation for Named Entity Recognition in English","authors":["O Agarwal, A Nenkova"],"snippet":"… We use two input word representations: the 300-d GloVe vectors trained on Common Crawl (Pennington et al., 2014), which is the dominant representation in NER, and the 1024-d contextual ELMo (Peters et al …","url":["https://aclanthology.org/2021.findings-acl.349.pdf"]}
{"year":"2021","title":"The Volctrans Neural Speech Translation System for IWSLT 2021","authors":["C Zhao, Z Liu, J Tong, T Wang, M Wang, R Ye, Q Dong… - arXiv preprint arXiv …, 2021"],"snippet":"… Data Combination and Sampling Strategy We train transformer models with different combina5http://www.statmt.org/wmt20/ translation-task.html, including Common Crawl, tions of data sets because increasing the model's diversity can benefit the model ensemble …","url":["https://arxiv.org/pdf/2105.07319"]}
{"year":"2021","title":"The Web Is Your Oyster--Knowledge-Intensive NLP against a Very Large Web Corpus","authors":["A Piktus, F Petroni, V Karpukhin, D Okhonko… - arXiv preprint arXiv …, 2021"],"snippet":"… CCNet processes Common Crawl by performing deduplication, language identification and quality filtering (articles are split into three quality tiers: head, … We pick the CCNet snapshot corresponding to the August 2019 Common Crawl …","url":["https://arxiv.org/pdf/2112.09924"]}
{"year":"2021","title":"TheEyeCorpus: Experiments in Reducing NLP Bias and Identifiability for Large LMs","authors":["J Herrera, D Bernal"],"snippet":"… 2020), a state of the art large language model, that used terabytes of the mostly unfiltered Common Crawl. Making corpuses is usually as simple as downloading and extracting the common crawl database. This isn’t the most ideal approach …","url":["https://openreview.net/pdf?id=CxMuhSuXIW"]}
{"year":"2021","title":"Thor at Touché 2021: Argument Retrieval for Comparative Questions","authors":["E Shirshakova, A Wattar - Working Notes of CLEF, 2021"],"snippet":"… doi:10.1145/2348283.2348429. [4] J. Bevendorff, B. Stein, M. Hagen, M. Potthast, Elastic ChatNoir: Search Engine for the ClueWeb and the Common Crawl, in: L. Azzopardi, A. Hanbury, G. Pasi, B. Piwowarski (Eds.), Advances in Information Retrieval …","url":["http://ceur-ws.org/Vol-2936/paper-219.pdf"]}
{"year":"2021","title":"Three Essays on the Role of Unstructured Data in Marketing Research","authors":["ISK Chakraborty - 2021"],"snippet":"Page 1. Abstract Three Essays on the Role of Unstructured Data in Marketing Research Ishita Sunity Kumar Chakraborty 2021 This thesis studies the use of firm and user-generated unstructured data (eg, text and videos) for …","url":["https://search.proquest.com/openview/c265768db80d44b3e2c4fef8b9b0df95/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Time and Cost Prediction Models for Language Classification Over a Large Corpus on Spark","authors":["JB Rodrigues, GC Vasconcelos, PRM Maciel - 2020 IEEE Symposium Series on …, 2020"],"snippet":"… All nodes were managed by a Debian GNU/Linux 9 OS, Apache Spark 2.3, and Apache Hadoop 2.9. 2 The Common Crawl Foundation - http://commoncrawl.org/ 1) Analysis of Variance: Fig. 1 displays the results over the log-transformed data …","url":["https://ieeexplore.ieee.org/abstract/document/9308299/"]}
{"year":"2021","title":"TMEKU System for the WAT2021 Multimodal Translation Task","authors":["Y Zhao, M Komachi, T Kajiwara, C Chu - red"],"snippet":"… Subsequently, we utilize fastText (Bojanowski et al., 2017) to learn subword embeddings. We use a pre-trained model4 containing two million word vectors trained with subword information on Common Crawl (600B tokens) …","url":["https://aclanthology.org/2021.wat-1.20.pdf"]}
{"year":"2021","title":"Token-Level Multilingual Epidemic Dataset for Event Extraction","authors":["S Mutuvi, E Boros, A Doucet, G Lejeune, A Jatowt… - International Conference on …, 2021"],"snippet":"… 8. XLM-RoBERTa-base was trained on 2.5TB of CommonCrawl data in 100 languages. References. 1. Aiello, AE, Renson, A., Zivich, PN: Social media-and internet-based disease surveillance for public health. Annu. Rev. Public …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86324-1_6"]}
{"year":"2021","title":"Touché Task 2: Comparative Argument Retrieval. A Document-based Search Engine for Answering Comparative Questions","authors":["D Helmrich, D Streitmatter, F Fuchs, M Heykeroth - Working Notes of CLEF, 2021"],"snippet":"… eg faster). The results are sentences from the Common Crawl2 retrieved via ElasticSearch [7] and then reduced to comparative sentences (either by certain keywords or a machine learning approach) and ranked. The system …","url":["http://ceur-ws.org/Vol-2936/paper-214.pdf"]}
{"year":"2021","title":"Toward Gender-Inclusive Coreference Resolution: An Analysis of Gender and Bias throughout the Machine Learning Lifecyle","authors":["YT Cao, H Daumé - Computational Linguistics, 2021"],"snippet":"Page 1. Toward Gender-Inclusive Coreference Resolution An Analysis of Gender and Bias throughout the Machine Learning Lifecyle∗ Yang Trista Cao∗∗ University of Maryland Hal Daumé III† Microsoft Research University of Maryland …","url":["https://direct.mit.edu/coli/article-pdf/doi/10.1162/coli_a_00413/1928379/coli_a_00413.pdf"]}
{"year":"2021","title":"Toward the Understanding of Deep Text Matching Models for Information Retrieval","authors":["L Chen, Y Lan, L Pang, J Guo, X Cheng - arXiv preprint arXiv:2108.07081, 2021"],"snippet":"… For the input word embeddings, we initialize the embedding layer with the 300-dimensional Glove[29] word vectors pre-trained in the 840B Common Crawl corpus4. For the out-of … 3https://github.com/NTMC-Community …","url":["https://arxiv.org/pdf/2108.07081"]}
{"year":"2021","title":"Towards a Machine Learning Based Generalizable Framework for Detecting COVID-19 Misinformation on Social Media","authors":["Y Chen - 2021"],"snippet":"Page 1. University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln …","url":["https://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1232&context=computerscidiss"]}
{"year":"2021","title":"TOWARDS A ROMANIAN END-TO-END AUTOMATIC SPEECH RECOGNITION BASED ON DEEPSPEECH2","authors":["AM AVRAM, P Vasile, D Tufiș"],"snippet":"… [17]) layers that have 768 units. The output of the LSTM layers is 7 https://commoncrawl org/ 8 We used langdetect to detect Romanian sentences: https://github.com/Mimino666/ langdetect. Page 4. Andrei-Marius AVRAM, Vasile PĂIȘ, Dan TUFIȘ 398 …","url":["https://academiaromana.ro/sectii2002/proceedings/doc2020-4/11-Avram_Tufis.pdf"]}
{"year":"2021","title":"Towards better subtitles: A multilingual approach for punctuation restoration of speech transcripts","authors":["NM Guerreiro, R Rei, F Batista - Expert Systems with Applications, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0957417421011180"]}
{"year":"2021","title":"Towards Change Detection in Privacy Policies with Natural Language Processing","authors":["A Adhikari, R Dewri - 2021 18th International Conference on Privacy …, 2021"],"snippet":"Privacy policies notify users about the privacy practices of websites, mobile apps, and other products and services. However, users rarely read them and struggle to understand their contents. Due to the complicated nature of these documents, it gets …","url":["https://ieeexplore.ieee.org/abstract/document/9647767/"]}
{"year":"2021","title":"Towards Continual Knowledge Learning of Language Models","authors":["J Jang, S Ye, S Yang, J Shin, J Han, G Kim, SJ Choi… - arXiv preprint arXiv …, 2021"],"snippet":"… 3CC-RECENTNEWS consists of 221,779 articles (∼168M tokens), which is estimated to be about 750 times smaller than C4, a cleansed version of the April 2019 Common Crawl dataset (https://commoncrawl.org/) that was …","url":["https://arxiv.org/pdf/2110.03215"]}
{"year":"2021","title":"Towards Continual Learning for Multilingual Machine Translation via Vocabulary Substitution","authors":["X Garcia, N Constant, AP Parikh, O Firat - arXiv preprint arXiv:2103.06799, 2021"],"snippet":"Page 1. Towards Continual Learning for Multilingual Machine Translation via Vocabulary Substitution Xavier Garcia and Noah Constant and Ankur P. Parikh and Orhan Firat Google Research Mountain View California …","url":["https://arxiv.org/pdf/2103.06799"]}
{"year":"2021","title":"Towards Controllable Neural Generation of Arguments","authors":["X Hua - 2021"],"snippet":"Page 1. Page 2. TOWARDS CONTROLLABLE NEURAL GENERATION OF ARGUMENTS A Dissertation Presented to The Khoury College of Computer Sciences of Northeastern University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy …","url":["https://search.proquest.com/openview/b002097758dccea01c78e5461afe631f/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Towards Generalization in Dialog through Inductive Biases","authors":["S Mehri - 2021"],"snippet":"Abstract Generalization is imperative in dialog research. Data-driven models have been shown to be capable of performing specific tasks in constrained contexts, given ample data. However, the complexity of human communication necessitates …","url":["http://shikib.com/proposal.pdf"]}
{"year":"2021","title":"Towards Learning a Joint Representation from Transformer in Multimodal Emotion Recognition","authors":["JJ Deng, CHC Leung - International Conference on Brain Informatics, 2021"],"snippet":"… modality of speech audio. As for text representation, we adopt fine-tuned Text-To-Text Transfer Transformer (T5) [18] model trained by a common crawl (C4) dataset to extract text embeddings. Operations of transfer learning …","url":["https://link.springer.com/chapter/10.1007/978-3-030-86993-9_17"]}
{"year":"2021","title":"Towards Learning Language Agnostic Features for NLP in Low-resource Languages","authors":["AJ Antony - 2020"],"snippet":"Page 1. Towards Learning Language Agnostic Features for NLP in Low-resource Languages Thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer Science and …","url":["http://web2py.iiit.ac.in/research_centres/publications/download/mastersthesis.pdf.bfe7226242875ee5.5468657369735f416c6c656e2e706466.pdf"]}
{"year":"2021","title":"Towards Learning Terminological Concept Systems from Multilingual Natural Language Text","authors":["L Wachowiak, C Lang, B Heinisch, D Gromann - 3rd Conference on Language, Data …, 2021"],"snippet":"… [6]. We further added to the data by manually annotating around 200 acronym-term pairs from ACTER with relations other than synonymy and another 271 sample sentences from the Common Crawl News Corpus5 … 5 …","url":["https://drops.dagstuhl.de/opus/volltexte/2021/14558/pdf/OASIcs-LDK-2021-22.pdf"]}
{"year":"2021","title":"Towards Making the Most of Dialogue Characteristics for Neural Chat Translation","authors":["Y Liang, C Zhou, F Meng, J Xu, Y Chen, J Su, J Zhou - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Towards Making the Most of Dialogue Characteristics for Neural Chat Translation Yunlong Liang1∗, Chulun Zhou2∗, Fandong Meng3, Jinan Xu1†, Yufeng Chen1, Jinsong Su2 and Jie Zhou3 1Beijing Key Lab of Traffic …","url":["https://arxiv.org/pdf/2109.00668"]}
{"year":"2021","title":"Towards Making the Most of Multilingual Pretraining for Zero-Shot Neural Machine Translation","authors":["G Chen, S Ma, Y Chen, D Zhang, J Pan, W Wang… - arXiv preprint arXiv …, 2021"],"snippet":"… 2018, SPM) learned on the full Common Crawl data that includes 250k subword tokens. We do not apply additional preprocessing, such as truecasing or normalizing punctuation/characters. Following XLMR, we add the [BOS] and [EOS] tokens at the …","url":["https://arxiv.org/pdf/2110.08547"]}
{"year":"2021","title":"Towards Model Understanding","authors":["D Pruthi - 2021"],"snippet":"While deep learning models have become increasingly accurate over the last decade, concerns about their (lack of) interpretability have taken a center stage. In response, a growing sub-field on interpretability and analysis of these models has …","url":["https://www.cs.cmu.edu/~ddanish/files/thesis.pdf"]}
{"year":"2021","title":"Towards More Effective and Economic Sparsely-Activated Model","authors":["H Jiang, K Zhan, J Qu, Y Wu, Z Fei, X Zhang, L Chen… - arXiv preprint arXiv …, 2021","L Chen, Z Dou, X Qiu, Z Guo, R Lai, J Wu, E Hu…"],"snippet":"… RealNews is a large scale dataset about news articles from Common Crawl OpenWebText dataset contains a large amount of text data from social media platform Reddit. They scraped all outbound links from Reddit as data sources and …","url":["https://ar5iv.labs.arxiv.org/html/2110.07431","https://arxiv.org/pdf/2110.07431"]}
{"year":"2021","title":"Towards Multi-modal Entity Resolution for Product Matching","authors":["M Wilke, E Rahm"],"snippet":"… Finally a procedure to query the internet archive2 for missing images is used. Due to its slow speed this method is only applied selectively to achieve higher coverage of images for the experiments. 1http://commoncrawl.org/ 2https://archive.org/ Page 3 …","url":["https://dbs.uni-leipzig.de/file/GvDB21_multi_modal_product_matching.pdf"]}
{"year":"2021","title":"Towards Reinforcement Learning for Pivot-based Neural Machine Translation with Non-autoregressive Transformer","authors":["E Tokarchuk, J Rosendahl, W Wang, P Petrushkov… - arXiv preprint arXiv …, 2021"],"snippet":"Page 1. Towards Reinforcement Learning for Pivot-based Neural Machine Translation with Non-autoregressive Transformer Evgeniia Tokarchuk 1 2 Jan Rosendahl 2 Weiyue Wang 2 Pavel Petrushkov 1 Tomer Lancewicki 1 Shahram Khadivi 1 Hermann Ney 2 …","url":["https://arxiv.org/pdf/2109.13097"]}
{"year":"2021","title":"Towards Selecting Informative Content for Cyber Threat Intelligence","authors":["P Panagiotou, C Iliou, K Apostolou, T Tsikrika… - 2021 IEEE International …, 2021"],"snippet":"… NER model. For their implementation, we used the Delft tool3, and the Glove Common Crawl embeddings4. The Bi- directional LSTM-CNNs-CRF architecture of [29] performs the best according to all metrics considered. We …","url":["https://ieeexplore.ieee.org/abstract/document/9527909/"]}
{"year":"2021","title":"Towards Target-dependent Sentiment Classification in News Articles","authors":["F Hamborg, K Donnay, B Gipp"],"snippet":"… category-dataset Page 10. 10 F. Hamborg et al. 23. Nagel, S.: Common Crawl: News Crawl (2016), https://web. archive.org/web/20191118111519/https://commoncrawl org/2016/10/ news-dataset-available/ 24. Nakov, P., Ritter …","url":["https://www.gipp.com/wp-content/papercite-data/pdf/hamborg2021.pdf"]}
{"year":"2021","title":"Towards the smart use of embedding and instance features for property matching","authors":["D Ayala, I Hernández, D Ruiz, E Rahm - 2021 IEEE 37th International Conference on …, 2021"],"snippet":"… To compute embeddings, we use the pre-trained GloVe approach [25]1, specifically for the uncased Common Crawl corpus that includes 300-dimensional vectors for 1.9 million words. Unknown words are mapped to a vector filled with zeroes …","url":["https://ieeexplore.ieee.org/abstract/document/9458651/"]}
{"year":"2021","title":"Towards Zero-Shot Cross-Lingual Named Entity Disambiguation","authors":["A Barrena Madinabeitia, A Soroa Echave… - 2021","A Barrena, A Soroa, E Agirre - Expert Systems with Applications, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://addi.ehu.es/bitstream/handle/10810/54482/1-s2.0-S0957417421009490-main.pdf?sequence=1&isAllowed=y","https://www.sciencedirect.com/science/article/pii/S0957417421009490"]}
{"year":"2021","title":"Traceability recovery between bug reports and test cases-a Mozilla Firefox case study","authors":["G Gadelha, F Ramalho, T Massoni - Automated Software Engineering, 2021"],"snippet":"Automatic recovery of traceability between software artifacts may promote early detection of issues and better calculate change impact. Information Retriev.","url":["https://link.springer.com/article/10.1007/s10515-021-00287-w"]}
{"year":"2021","title":"Trafilatura: A Web Scraping Library and Command-Line Tool for Text Discovery and Extraction","authors":["A Barbaresi - 2021"],"snippet":"… Data collection approaches using the CommonCrawl1 have flourished as they allow for faster download and processing by skipping (or more precisely outsourcing) the crawling phase … 1https://commoncrawl.org Page 2. 123 …","url":["https://aclanthology.org/2021.acl-demo.15.pdf"]}
{"year":"2021","title":"Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation","authors":["O Press, NA Smith, M Lewis - arXiv preprint arXiv:2108.12409, 2021"],"snippet":"Page 1. TRAIN SHORT, TEST LONG: ATTENTION WITH LINEAR BIASES ENABLES INPUT LENGTH EXTRAPOLATION Ofir Press1,2 Noah A. Smith1,3 Mike Lewis2 1Paul G. Allen School of Computer Science & Engineering …","url":["https://arxiv.org/pdf/2108.12409"]}
{"year":"2021","title":"Training Domain Specific Multilingually Aligned Word Embeddings","authors":["X Wang, C Fan, M Frantzen - 2021"],"snippet":"… FastText The pre-trained Fasttext embeddings follow the training procedure proposed in [5]. They use a mixture of Wikipedia and CommonCrawl as training data and learn 300 dimensional embeddings with a CBOW model defined …","url":["https://xinpeng-wang.github.io/pdfs/nlp_final_report.pdf"]}
{"year":"2021","title":"Training ELECTRA Augmented with Multi-word Selection","authors":["J Shen, J Liu, T Liu, C Yu, J Han - arXiv preprint arXiv:2106.00139, 2021"],"snippet":"Page 1. Training ELECTRA Augmented with Multi-word Selection Jiaming Shen∇⋆, Jialu Liu O , Tianqi Liu O , Cong Yu O , Jiawei Han∇ ∇University of Illinois Urbana-Champaign, IL, USA, 3 Google Research, NY, USA ∇{js2 …","url":["https://arxiv.org/pdf/2106.00139"]}
{"year":"2021","title":"Transfer Learning for Automated Responses to the BDI Questionnaire","authors":["C Spartalis, G Drosatos, A Arampatzis - Working Notes of CLEF, 2021"],"snippet":"… batches, and over more data. More specifically, its pretraining corpus also includes CC-News (portion of the CommonCrawl News dataset [44]), OpenWebText [45], and Stories [46]). Furthermore, there are some modifications …","url":["http://ceur-ws.org/Vol-2936/paper-84.pdf"]}
{"year":"2021","title":"Transfer Learning for Sequence Generation: from Single-source to Multi-source","authors":["X Huang, J Xu, M Sun, Y Liu - arXiv preprint arXiv:2105.14809, 2021"],"snippet":"Page 1. Transfer Learning for Sequence Generation: from Single-source to Multi-source Xuancheng Huang1, Jingfang Xu4, Maosong Sun1,3, and Yang Liu1,2,3∗ 1Dept. of Comp. Sci. & Tech., BNRist Center, Institute for AI, Tsinghua …","url":["https://arxiv.org/pdf/2105.14809"]}
{"year":"2021","title":"Transfer Learning Model for Disrupting Misinformation During a COVID-19 Pandemic","authors":["R Raju, S Bhandari, SA Mohamud, EN Ceesay - 2021 IEEE 11th Annual Computing …, 2021"],"snippet":"… 19–27, IEEE Computer Society, 2015. [30] S. Nagel, “Cc-news.” https://commoncrawl.org/ 2016/10/news-datasetavailable/, 2016. [31] A. Gokaslan and V. Cohen, “Openwebtext corpus.” https://skylion007.github.io/OpenWebTextCorpus/, 2016 …","url":["https://ieeexplore.ieee.org/abstract/document/9376066/"]}
{"year":"2021","title":"Transfer Learning Techniques for Sequence Labeling in Network File System Specifications","authors":["A Singh - 2021"],"snippet":"… English Wikipedia and BookCorpus (Zhu et al., 2015), and RoBERTa (Liu et al., 2019) is pre-trained on Common-Crawl News (Nagel, 2016), OpenWebText (Radford et al., 2019), and Stories (Trinh and Le, 2019) …","url":["https://search.proquest.com/openview/754b6463cd54e1b58537a7d6558d4825/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Transformer-Based Automatic Punctuation Prediction and Word Casing Reconstruction of the ASR Output","authors":["J Švec, J Lehečka, L Šmídl, P Ircing - International Conference on Text, Speech, and …, 2021"],"snippet":"… For experiments with English data, we used the Google's T5-base English model (version from October 21th, 2019) 3 trained from Common Crawl data 4 . We replicated the same pre-processing procedure to obtain the …","url":["https://link.springer.com/chapter/10.1007/978-3-030-83527-9_7"]}
{"year":"2021","title":"Transformer-based identification of stochastic information cascades in social networks using text and image similarity","authors":["P Kasnesis, R Heartfield, X Liang, L Toumanidis… - Applied Soft Computing, 2021"],"snippet":"… An advantage of RoBERTa against BERT for text-based information cascade identification is its pretrained architecture which benefits from a more diverse range of datasets (larger corpus). For example, its training corpus includes the CommonCrawl News dataset.","url":["https://www.sciencedirect.com/science/article/pii/S1568494621003367"]}
{"year":"2021","title":"TRANSFORMERS AND THEIR APPLICATIONS IN NATURAL LANGUAGE PROCESSING","authors":["N Sarzhan"],"snippet":"Page 1. Page 2. E¨OTV¨OS LORÁND UNIVERSITY FACULTY OF SCIENCE TRANSFORMERS AND THEIR APPLICATIONS IN NATURAL LANGUAGE PROCESSING MASTER'S THESIS Nurzhan Sarzhan MSc mathematics …","url":["https://web.cs.elte.hu/blobs/diplomamunkak/msc_alkmat/2021/nurzhan_sarzhan.pdf"]}
{"year":"2021","title":"Transforming Term Extraction: Transformer-Based Approaches to Multilingual Term Extraction Across Domains","authors":["C Lang, L Wachowiak, B Heinisch, D Gromann"],"snippet":"… XLM-R (Conneau et al., 2020) is a multilingual variant of BERT, which was pretrained in 100 languages using 2.5 terabytes of Common Crawl data. Moreover, it makes use of the improved training routine introduced by RoBERTa (Liu et al., 2019) …","url":["https://aclanthology.org/2021.findings-acl.316.pdf"]}
{"year":"2021","title":"Translate and Classify: Improving Sequence Level Classification for English-Hindi Code-Mixed Data","authors":["D Gautam, K Gupta, M Shrivastava - Proceedings of the Fifth Workshop on …, 2021"],"snippet":"… auto-encoder. It has been pre-trained using the BART (Lewis et al., 2020) objective on large-scale monolingual corpora of 25 languages extracted from Common Crawl1 (Wenzek et al., 2020; Conneau et al., 2020). Both English …","url":["https://www.aclweb.org/anthology/2021.calcs-1.3.pdf"]}
{"year":"2021","title":"TrollsWithOpinion: A Dataset for Predicting Domain-specific Opinion Manipulation in Troll Memes","authors":["S Suryawanshi, BR Chakravarthi, M Arcan, S Little… - arXiv preprint arXiv …, 2021"],"snippet":"… 5.2 Deep Learning baselines Bag of words (BOW) : This unimodal baseline used text representation in the form of 300-dimensional pre-trained GloVe Pennington et al. (2014) (common crawl) word embeddings, which were fed to a classification layer …","url":["https://arxiv.org/pdf/2109.03571"]}
{"year":"2021","title":"TunBERT: Pretrained Contextualized Text Representation for Tunisian Dialect","authors":["A Messaoudi, A Cheikhrouhou, H Haddad, N Ferchichi… - arXiv preprint arXiv …, 2021"],"snippet":"… They also concluded that a model trained on a CommonCrawl-based corpus performed consistently better than the one trained on the French Wikipedia. They suggested that a 4 GB heterogeneous dataset in terms of genre and style is large enough as a …","url":["https://arxiv.org/pdf/2111.13138"]}
{"year":"2021","title":"Turkic Interlingua: A Case Study of Machine Translation in Low-resource Languages","authors":["J Mirzakhalov - 2021"],"snippet":"… production. Table 4 shows results from the CCAligned dataset, which is a parallel dataset compiled from 68 Common Crawl snapshots … shown in Table 6. Data for OSCAR is also a compilation of Common Crawl snapshots and follows a similar filtering process …","url":["https://scholarcommons.usf.edu/cgi/viewcontent.cgi?article=10026&context=etd"]}
{"year":"2021","title":"Twin Question Pair Classification","authors":["A Sharma, SS Jha, S Arora, S Garg, S Tayal - Smart and Sustainable Intelligent …, 2021"],"snippet":"… 4 Embedding Embedding maps words in a text document to number space (these number-space representations are called word vectors). We have to use Fast text embedding [6]. The Fast text model for English is pre-trained …","url":["http://books.google.de/books?hl=en&lr=lang_en&id=7fEhEAAAQBAJ&oi=fnd&pg=PA215&dq=commoncrawl&ots=oxs_QNhHWM&sig=OazT1mzKhUG_fOXx1LjGsEMtRKI"]}
{"year":"2021","title":"Twitter Dataset and Evaluation of Transformers for Turkish Sentiment Analysis","authors":["A Köksal, A Özgür - 2021 29th Signal Processing and Communications …, 2021"],"snippet":"… Kelime Büyüklü˘gü 128,000 119,547 250,002 Kullanılan Diller Türkçe Türkçe + 103 dil Türkçe + 99 dil E˘gitim Derlemi Oscar + Wikipedia + Opus Wikipedia CommonCrawl vektör uzunlu˘gu, dikkat ve gizli katmanların sayısı bu üç mo- delde de aynıdır …","url":["https://ieeexplore.ieee.org/abstract/document/9477814/"]}
{"year":"2021","title":"Two Parents, One Child: Dual Transfer for Low-Resource Neural Machine Translation","authors":["M Zhang, L Li, Q Liu"],"snippet":"Page 1. Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 2726–2738 August 1–6, 2021. ©2021 Association for Computational Linguistics 2726 Two Parents, One Child: Dual Transfer for Low-Resource Neural Machine Translation …","url":["https://aclanthology.org/2021.findings-acl.241.pdf"]}
{"year":"2021","title":"Two-stage Visual Cues Enhancement Network for Referring Image Segmentation","authors":["Y Jiao, Z Jie, W Luo, J Chen, YG Jiang, X Wei, L Ma - arXiv preprint arXiv:2110.04435, 2021"],"snippet":"… As for language processing, GloVe [27] word embeddings pretrained on Common Crawl 840B tokens are firstly used for constructing word embedding look-up table. Then a LSTM takes GloVe word embeddings as input for capturing context information …","url":["https://arxiv.org/pdf/2110.04435"]}
{"year":"2021","title":"Type-and Token-based Word Embeddings in the Digital Humanities","authors":["A Ehrmanntraut, T Hagen, L Konle, F Jannidis - Proceedings http://ceur-ws. org ISSN, 2021"],"snippet":"Abstract In the general perception of the NLP community, the new dynamic, context-sensitive, token-based embeddings from language models like BERT have replaced the older static, type-based embeddings like word2vec or fastText, due to their better …","url":["http://ceur-ws.org/Vol-2989/long_paper35.pdf"]}
{"year":"2021","title":"UGTO: Uncommon Words and Proper Nouns","authors":["X Zhong, E Cambria - Time Expression and Named Entity Recognition, 2021"],"snippet":"… they are (1) word2vec, which is trained on the Google News dataset [13], (2) GloVe, which is trained on the Wikipedia 2014 and Gigaword 5 corpora [20], and (3) FastText, which is trained on the Wikipedia 2017, UMBC corpus …","url":["https://link.springer.com/chapter/10.1007/978-3-030-78961-9_6"]}
{"year":"2021","title":"Uncovering the Limits of Text-based Emotion Detection","authors":["N Alvarez-Gonzalez, A Kaltenbrunner, V Gómez - arXiv preprint arXiv:2109.01900, 2021"],"snippet":"Page 1. Uncovering the Limits of Text-based Emotion Detection Nurudin Alvarez-Gonzalez ∗, Andreas Kaltenbrunner, Vicenç Gómez nuralgon@gmail.com, andreas.kaltenbrunner@ upf.edu, vicen.gomez@upf.edu Universitat Pompeu Fabra. Barcelona, Spain. Abstract …","url":["https://arxiv.org/pdf/2109.01900"]}
{"year":"2021","title":"Understanding and Countering Stereotypes: A Computational Approach to the Stereotype Content Model","authors":["KC Fraser, I Nejadgholi, S Kiritchenko"],"snippet":"… trained word em- beddings in the gensim library ( ˇRehurek and Sojka, 2010), with the results given in Table 2. The FastText embeddings generally outperform the other embeddings on this task, with the 2M word model trained …","url":["https://www.svkir.com/papers/Fraser-et-al-SCM-ACL2021.pdf"]}
{"year":"2021","title":"Understanding the Effect of Out-of-distribution Examples and Interactive Explanations on Human-AI Decision Making","authors":["H Liu, V Lai, C Tan - arXiv preprint arXiv:2101.05303, 2021"],"snippet":"… Race. There are 7,214 defendants in this dataset. • BIOS [9]. This dataset contains hundreds of thousands of online biographies from the Common Crawl corpus. The task is to predict a person's profession given a biography …","url":["https://arxiv.org/pdf/2101.05303"]}
{"year":"2021","title":"Ungoliant: An optimized pipeline for the generation of a very large-scale multilingual web corpus","authors":["J Abadji, PJ Ortiz Suárez, L Romary, B Sagot - … of the Workshop on Challenges in the …, 2021"],"snippet":"… The most recent Common Crawl dump contains 64,000 shards. Each shard is composed of numerous records, and each record holds textual content along with metadata. While CommonCrawl shards hold document-level …","url":["https://ids-pub.bsz-bw.de/files/10468/Abadji_Suarez_Romary_Ungoliant_2021.pdf"]}
{"year":"2021","title":"Unified Open-Domain Question Answering with Structured and Unstructured Knowledge","authors":["B Oguz, X Chen, V Karpukhin, S Peshterliev… - arXiv preprint arXiv …, 2020"],"snippet":"… KBQA is ultimately limited in its coverage of facts and the types of questions it can answer. On the other hand, large collections of text such as Wikipedia or CommonCrawl promise to be a richer source of knowledge for truly open do- main question answering systems …","url":["https://arxiv.org/pdf/2012.14610"]}
{"year":"2021","title":"Universal Sentence Representation Learning with Conditional Masked Language Model","authors":["Z Yang, Y Yang, D Cer, J Law, E Darve - arXiv preprint arXiv:2012.14388, 2020"],"snippet":"… For training English sentence encoders with CMLM, we use three Common Crawl dumps … Similarly, we also train QuickThought, a competitive unsupervised sentence representations learning model, on the same …","url":["https://arxiv.org/pdf/2012.14388"]}
{"year":"2021","title":"University of Copenhagen Participation in TREC Health Misinformation Track 2020","authors":["LC Lima, DB Wright, I Augenstein, M Maistro - arXiv preprint arXiv:2103.02462, 2021"],"snippet":"… The corpus of the Total Recall and Ad Hoc Tasks consits of news articles in the CommonCrawl News crawl, sampled from January, 1st 2020 … Since there are no credibility assessments for the Common Crawl News sample …","url":["https://arxiv.org/pdf/2103.02462"]}
{"year":"2021","title":"University of Regensburg@ PAN: Profiling Hate Speech Spreaders on Twitter","authors":["KO Akomeah, U Kruschwitz, B Ludwig - CLEF, 2021"],"snippet":"… This preprocessor is a companion to the BERT models for preprocsssing plain text inputs into the input format expected by BERT. The model uses a vocabulary for multilingual models extracted from Wikipedia, CommonCrawl, and translation pairs from the Web …","url":["http://ceur-ws.org/Vol-2936/paper-184.pdf"]}
{"year":"2021","title":"Unraveling Social Perceptions & Behaviors towards Migrants on Twitter","authors":["A Khatua, W Nejdl - arXiv preprint arXiv:2112.06642, 2021"],"snippet":"We draw insights from the social psychology literature to identify two facets of Twitter deliberations about migrants, ie, perceptions about migrants and behaviors towards mi-grants. Our theoretical anchoring helped us in identifying two prevailing …","url":["https://arxiv.org/pdf/2112.06642"]}
{"year":"2021","title":"Unsupervised Clustering of Structured and Unstructured Text Collections","authors":["L Bungum - 2021"],"snippet":"Page 1. Lars B u n g u m Doctoral theses at NTNU, 202 1 :179 ISBN 978-82-326-5390-4 (printed ver.) ISBN 978-82-326-6966-0 (electronic ver.) ISSN 1503-8181 (printed ver.) ISSN 2703-8084 (electronic ver.) Do cto ral …","url":["https://ntnuopen.ntnu.no/ntnu-xmlui/bitstream/handle/11250/2759145/Lars%20Bungum_PhD.pdf?sequence=1"]}
{"year":"2021","title":"Unsupervised Cross-Domain and Cross-Lingual Methods for Text Classification, Slot-Filling, and Question-Answering","authors":["J Krishnan - 2021"],"snippet":"Page 1. UNSUPERVISED CROSS-DOMAIN AND CROSS-LINGUAL METHODS FOR TEXT CLASSIFICATION, SLOT-FILLING, AND QUESTION-ANSWERING by Jitin Krishnan A Dissertation Submitted to the Graduate Faculty …","url":["https://search.proquest.com/openview/5ef2690ad3bc61870ee72a9048d3a4f9/1?pq-origsite=gscholar&cbl=18750&diss=y"]}
{"year":"2021","title":"Unsupervised Neural Machine Translation with Generative Language Models Only","authors":["JM Han, I Babuschkin, H Edwards, A Neelakantan… - arXiv preprint arXiv …, 2021"],"snippet":"We show how to derive state-of-the-art unsupervised neural machine translation systems from generatively pre-trained language models. Our method consists of three steps: few-shot amplification, distillation, and backtranslation. We first use the …","url":["https://arxiv.org/pdf/2110.05448"]}
{"year":"2021","title":"Uppsala NLP at SemEval-2021 Task 2: Multilingual Language Models for Fine-tuning and Feature Extraction in Word-in-Context Disambiguation","authors":["H You, X Zhu, S Stymne - arXiv preprint arXiv:2104.03767, 2021"],"snippet":"… 3 Multilingual Language Models 3.1 XLMR XLMR (XLM-RoBERTa) is a scaled cross-lingual sentence encoder (Conneau et al., 2020), which is trained on 2.5T of data obtained from Common Crawl that covers more than 100 languages …","url":["https://arxiv.org/pdf/2104.03767"]}
{"year":"2021","title":"UPV at TREC Health Misinformation Track 2021 Ranking with SBERT and Quality Estimators","authors":["IB Schlicht, AFM de Paula, P Rosso - arXiv preprint arXiv:2112.06080, 2021"],"snippet":"Health misinformation on search engines is a significant problem that could negatively affect individuals or public health. To mitigate the problem, TREC organizes a health misinformation track. This paper presents our submissions to this …","url":["https://arxiv.org/pdf/2112.06080"]}
{"year":"2021","title":"UQuAD1. 0: Development of an Urdu Question Answering Training Data for Machine Reading Comprehension","authors":["S Kazi, S Khoja - arXiv preprint arXiv:2111.01543, 2021"],"snippet":"… The Third model used is XLM-Roberta, a multi-lingual pre-trained transformer model from common Crawl trained on 100 languages, including Urdu. XLM-Roberta outperformed previous multi-lingual models such as mBERT and XLM on a variety of …","url":["https://arxiv.org/pdf/2111.01543"]}
{"year":"2021","title":"User Identification in Online Social Networks using Graph Transformer Networks","authors":["KNP Kumar, ML Gavrilova - 2021 18th International Conference on Privacy …, 2021"],"snippet":"… We used the model with 2 million word vectors trained with subword information on Common Crawl (600B tokens). (v) Skip-thought Vectors (StV) [30]: A Gated Recurrent Unit (GRU) encoder trained to predict adjacent sentences in a books …","url":["https://ieeexplore.ieee.org/abstract/document/9647749/"]}
{"year":"2021","title":"Using Embedding-Based Similarities to Improve Lexical Resources","authors":["NV Loukachevitch, MM Tikhomirov, EA Parkhomenko - Lobachevskii Journal of …, 2021"],"snippet":"In this paper we discuss the usefulness of applying semi-automatic checking procedures to existing thesauri for natural language processing—large man.","url":["https://link.springer.com/article/10.1134/S1995080221070167"]}
{"year":"2021","title":"Using mixed methods to study the historical use of web beacons in web tracking","authors":["J Nielsen - International Journal of Digital Humanities, 2021"],"snippet":"… pixel. The approach is inspired by the study of web beacons by Ruohonen and Leppänen (2018) and Schelter and Kunegis' study of third-party embeddings in the Common Crawl 2012 corpus (Schelter & Kunegis, 2016, 2018) …","url":["https://link.springer.com/article/10.1007/s42803-021-00033-4"]}
{"year":"2021","title":"USING SEMANTIC RELATIONSHIPS TO ENHANCE NEURAL WORD EMBEDDINGS","authors":["Y Yin, X Sun, H Lv, P Wang, H Ma, D Yang"],"snippet":"Neural language models have significantly improved current natural language understanding tasks. However, distributional semantics, derived from neural language models is less competitive in computing semantic relatedness or similarity …","url":["https://www.researchgate.net/profile/Dongqiangdongqiang-Yang/publication/356592344_USING_SEMANTIC_RELATIONSHIPS_TO_ENHANCE_NEURAL_WORD_EMBEDDINGS/links/61a2f3196a0f6e76c2b43389/USING-SEMANTIC-RELATIONSHIPS-TO-ENHANCE-NEURAL-WORD-EMBEDDINGS.pdf"]}
{"year":"2021","title":"Using Visual Features to Generate Language Context","authors":["I Zachmann - 2021"],"snippet":"Page 1. Using Visual Features to Generate Language Context Undergraduate Honors Research Thesis Presented in Partial Fulfillment of the Requirements for Undergraduate Honors Research Thesis in the College of Engineering at The Ohio State University By …","url":["https://kb.osu.edu/bitstream/handle/1811/92666/Thesis.pdf?sequence=1"]}
{"year":"2021","title":"Using Word Embeddings to Determine Concepts of Values In Insurance Claim Spreadsheets","authors":["V Justnes - 2021"],"snippet":"Page 1. Using Word Embeddings to Determine Concepts of Values In Insurance Claim Spreadsheets Vemund Justnes Thesis submitted for the degree of Master in Informatics: Programming and System Architecture 60 credits Institute for Informatics …","url":["https://www.duo.uio.no/bitstream/handle/10852/87784/1/using_word_embeddings_for_concept_determiniation.pdf"]}
{"year":"2021","title":"Using word embeddings to probe sentiment associations of politically loaded terms in news and opinion articles from news media outlets","authors":["D Rozado, M al-Gharbi - Journal of Computational Social Science, 2021"],"snippet":"… word embeddings also absorb many of the biases that are prevalent within a cultural system, as a result of embedding models being trained on culturally relevant language artifacts that are presumed to contain cultural biases, such …","url":["https://link.springer.com/article/10.1007/s42001-021-00130-y"]}
{"year":"2021","title":"Utility-Preserving Anonymization of Textual Documents","authors":["FAM Hassan - 2021"],"snippet":"Page 1. UTILITY-PRESERVING ANONYMIZATION OF TEXTUAL DOCUMENTS Fadi Abdulfattah Mohammed Hassan ADVERTIMENT. L'accés als continguts d'aquesta tesi doctoral i la seva utilització ha de respectar els drets de la persona autora …","url":["https://84.88.27.106/bitstream/handle/10803/672012/TESI%20Fadi%20Abdulfattah%20Mohammed%20Hassan.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Utility-Preserving Privacy Protection of Textual Documents via Word Embeddings","authors":["F Hassan, D Sanchez, J Domingo-Ferrer - IEEE Transactions on Knowledge and Data …, 2021"],"snippet":"Page 1. 1041-4347 (c) 2021 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This …","url":["https://ieeexplore.ieee.org/abstract/document/9419784/"]}
{"year":"2021","title":"UXLA: A Robust Unsupervised Data Augmentation Framework for Zero-Resource Cross-Lingual NLP","authors":["MS Bari, T Mohiuddin, S Joty"],"snippet":"… Manning, 2019). We use XLM-R masked LM (Conneau et al., 2020) as our vicinity model θmlm, which is trained on massive multilingual corpora (2.5 TB of Common-Crawl data in 100 languages). The Page 4. 1981 Algorithm 1 …","url":["https://aclanthology.org/2021.acl-long.154.pdf"]}
{"year":"2021","title":"UzBERT: pretraining a BERT model for Uzbek","authors":["B Mansurov, A Mansurov - arXiv preprint arXiv:2108.09814, 2021"],"snippet":"… Page 3. 3 Uzbek Wikipedia and the Common Crawl corpus. According … 2006], where full words are replaced by their abbreviations to save space in print. The filtered Common Crawl corpus5 used by XLM-R [Conneau et al. 2019 …","url":["https://arxiv.org/pdf/2108.09814"]}
{"year":"2021","title":"Validating Word Embedding as a Tool for the Psychological Sciences","authors":["V Swift - 2021"],"snippet":"Page 1. Validating Word Embedding as a Tool for the Psychological Sciences by Victor Swift A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy Department of Psychology University of …","url":["https://tspace.library.utoronto.ca/bitstream/1807/104927/3/Swift_Victor_202103_PhD_thesis.pdf"]}
{"year":"2021","title":"Variational Autoencoder with Interactive Attention for Affective Text Generation","authors":["R Chen, J Wang, X Zhang - CCF International Conference on Natural Language …, 2021"],"snippet":"… The word embeddings used in this experiment was pre-trained on Common Crawl 840B 1 by GloVe [10]. For baselines, the polarity label \\(\\mathbf {c}\\) was treated as a scalar (ie, 0 or 1) and directly concatenated into the latent variables. For …","url":["https://link.springer.com/chapter/10.1007/978-3-030-88483-3_9"]}
{"year":"2021","title":"VECO: Variable and Flexible Cross-lingual Pre-training for Language Understanding and Generation","authors":["F Luo, W Wang, J Liu, Y Liu, B Bi, S Huang, F Huang…"],"snippet":"… mBART (Liu et al., 2020b) Encoder-decoder 680M 12 12 25 250k CommonCrawl VECO Flexible 662M 24* 50 250k CommonCrawl + Translation … For monolingual training datasets, we reconstruct CommonCrawl Corpus used in XLM-R (Conneau et al., 2019) …","url":["https://aclanthology.org/2021.acl-long.308.pdf"]}
{"year":"2021","title":"Vera: Prediction Techniques for Reducing Harmful Misinformation in Consumer Health Search","authors":["R Pradeep, X Ma, R Nogueira, J Lin - Proceedings of the 44th Annual International …, 2021"],"snippet":"… The document collection comprises news articles from the CommonCrawl news crawl spanning the first four months of 2020 (January 1st, 2020 to April 30th, 2020), covering the onset of the COVID-19 pandemic. This corpus …","url":["https://cs.uwaterloo.ca/~jimmylin/publications/Pradeep_etal_SIGIR2021.pdf"]}
{"year":"2021","title":"VERB: Visualizing and Interpreting Bias Mitigation Techniques for Word Representations","authors":["A Rathore, S Dev, JM Phillips, V Srikumar, Y Zheng… - arXiv preprint arXiv …, 2021"],"snippet":"… It also provides a larger GLoVe embedding (300-dimensional with the 100K most frequent words) from the Common Crawl corpus (https://commoncrawl.org/). 5 USE CASES We demonstrate the efficacy of VERB on several examples …","url":["https://arxiv.org/pdf/2104.02797"]}
{"year":"2021","title":"VieSum: How Robust Are Transformer-based Models on Vietnamese Summarization?","authors":["H Nguyen, L Phan, J Anibal, A Peltekian, H Tran - arXiv preprint arXiv:2110.04257, 2021"],"snippet":"… Common Crawl is a publicly web archive that provides text scraped from websites. These datasets (C4, CC25, and Common Crawl) may include Vietnamese Wikipedia languages website that help the model excel in constrained domain corpus (Wikipedia) …","url":["https://arxiv.org/pdf/2110.04257"]}
{"year":"2021","title":"Visual and Affective Multimodal Models of Word Meaning in Language and Mind","authors":["S De Deyne, DJ Navarro, G Collell, A Perfors - Cognitive Science, 2021"],"snippet":"Abstract One of the main limitations of natural language‐based approaches to meaning is that they do not incorporate multimodal representations the way humans do. In this study, we evaluate how wel...","url":["https://onlinelibrary.wiley.com/doi/pdf/10.1111/cogs.12922"]}
{"year":"2021","title":"Visual Grounding Strategies for Text-Only Natural Language Processing","authors":["D Sileo - arXiv preprint arXiv:2103.13942, 2021"],"snippet":"… A linear projection maps the region features to the input space of the cross-modal encoder. We also provide K rank 6We use the CommonCrawl version https:// fasttext.cc/docs/en/english-vectors.html. and discard stop-words with NLTK …","url":["https://arxiv.org/pdf/2103.13942"]}
{"year":"2021","title":"Visualizing large-scale high-dimensional data via hierarchical embedding of KNN graphs","authors":["H Zhu, M Zhu, Y Feng, D Cai, Y Hu, S Wu, X Wu… - Visual Informatics, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S2468502X21000292"]}
{"year":"2021","title":"VITALITY: Promoting Serendipitous Discovery of Academic Literature with Transformers & Visual Analytics","authors":["A Narechania, A Karduni, R Wesslen, E Wall - arXiv preprint arXiv:2108.03366, 2021"],"snippet":"… Unlike many pre-trained language models that use a general corpus like Wikipedia or the Common Crawl [14, 43, 49], SPECTER was pre-trained on academic literature (sciBERT [?]) and fine-tuned with citations which …","url":["https://arxiv.org/pdf/2108.03366"]}
{"year":"2021","title":"VOH. CoLAB at TREC 2020 Health Misinformation Track⋆","authors":["SN Gonçalves, F Martins"],"snippet":"… 2.1 Collection The collection contains 65 million news articles from CommonCrawl News4 corresponding to the period from January to April of 2020 … Recently, vinegar has been promoted as a disinfectant (...) 4 …","url":["https://trec.nist.gov/pubs/trec29/papers/vohcolab.HM.pdf"]}
{"year":"2021","title":"Voted In, Standing Out: Public Response to Immigrants' Political Accession","authors":["G Grossman, S Zonszein - 2021"],"snippet":"… newspapers, covering the general elections from 2010–2019.8 This data is from Common Crawl, which is an open repository of web crawl data. We assume that an article refers to a candidate's ethnic group when three conditions are met: 1) the publication date is …","url":["https://files.osf.io/v1/resources/xd4wk/providers/osfstorage/614782978ae0920335d8c84c?action=download&direct&version=2"]}
{"year":"2021","title":"We Need to Talk About Data: The Importance of Data Readiness in Natural Language Processing","authors":["F Olsson, M Sahlgren - arXiv preprint arXiv:2110.05464, 2021"],"snippet":"In this paper, we identify the state of data as being an important reason for failure in applied Natural Language Processing (NLP) projects. We argue that there is a gap between academic research in NLP and its application to problems outside …","url":["https://arxiv.org/pdf/2110.05464"]}
{"year":"2021","title":"Web Archive Analytics","authors":["M Völske, J Bevendorff, J Kiesel, B Stein, M Fröbe… - INFORMATIK 2020, 2021"],"snippet":"… in Figure 5, beginning with the bottom-most data acquisition layerȷ Primary sources for data ingestion include web crawls and web archives, such as the aforementioned Internet Archive, the Common Crawl,13 the older … 13 …","url":["https://dl.gi.de/bitstream/handle/20.500.12116/34759/A8-1.pdf?sequence=1&isAllowed=y"]}
{"year":"2021","title":"Web Content Authentication: A Machine Learning Approach to Identify Fake and Authentic Web Pages on Internet","authors":["J Ashok, P Badoni - … Technology for Competitive Strategies (ICTCS 2020) …"],"snippet":"… SpringerLink (Springer, 21 June 2017). www. link. springer. com/chapter/10.1007/978- 3-319-69784-0_15 42. Link to GitHub where code is hosted. https://github. com/Jkrish1011/Web-Content-Authentic ator 43. https://commoncrawl. org/","url":["http://books.google.de/books?hl=en&lr=lang_en&id=Dwo3EAAAQBAJ&oi=fnd&pg=PA85&dq=commoncrawl&ots=RCTLIhR7zF&sig=_9zgxY48BSktz0Jk-c6Y7XTbVf0"]}
{"year":"2021","title":"Web Table Classification based on Visual Features","authors":["B Bühler, H Paulheim - arXiv preprint arXiv:2103.05110, 2021"],"snippet":"… Lehmberg et al. [17] conduct web table classification in order to create a large open corpus of relational HTML tables, ie, the WebDataCommons HTML Ta- bles Dataset3 based on the 2012 version of the Common Crawl web corpus …","url":["https://arxiv.org/pdf/2103.05110"]}
{"year":"2021","title":"WeChat Neural Machine Translation Systems for WMT21","authors":["P Li, J Xu, J Zhou","X Zeng, Y Liu, E Li, Q Ran, F Meng, P Li, J Xu, J Zhou - arXiv preprint arXiv …, 2021"],"snippet":"… 4.2 Dataset For each language pair, the bilingual training data is the combination of all parallel data released by WMT21. For monolingual data, we filter out clean data from News Crawl, Common Crawl and Ex- tended Common Crawl …","url":["https://ar5iv.labs.arxiv.org/html/2108.02401","https://arxiv.org/pdf/2108.02401"]}
{"year":"2021","title":"What Can Unsupervised Machine Translation Contribute to High-Resource Language Pairs?","authors":["K Marchisio, M Freitag, D Grangier - arXiv preprint arXiv:2106.15818, 2021"],"snippet":"… As bilingual training data, we use News Commentary v13, Europarl v7, Common Crawl and EU Press Release. We deduplicate and filter out pairs with > 250 to- kens in either language or length ratio over 1.5. This results in 5.2 million parallel sentence pairs …","url":["https://arxiv.org/pdf/2106.15818"]}
{"year":"2021","title":"What Does Your Bio Say? Inferring Twitter Users' Depression Status From Multimodal Profile Information Using Deep Learning","authors":["S Ghosh, A Ekbal, P Bhattacharyya - IEEE Transactions on Computational Social …, 2021"],"snippet":"… Word embedding has been trained on the Common Crawl corpus (840 billion tokens). Padding arrays of zeros are done to make input sequences the same length. In our training set, we create an embedding matrix (E) that contains the embedding …","url":["https://ieeexplore.ieee.org/abstract/document/9567734/"]}
{"year":"2021","title":"What is this cluster about? Explaining textual clusters by extracting relevant keywords","authors":["A Penta, A Pal - Knowledge-Based Systems, 2021"],"snippet":"JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page. Skip to main content Skip to article …","url":["https://www.sciencedirect.com/science/article/pii/S0950705121006043"]}
{"year":"2021","title":"What Makes Sentences Semantically Related: A Textual Relatedness Dataset and Empirical Study","authors":["M Abdalla, K Vishnubhotla, SM Mohammad - arXiv preprint arXiv:2110.04845, 2021"],"snippet":"The degree of semantic relatedness (or, closeness in meaning) of two units of language has long been considered fundamental to understanding meaning. Automatically determining relatedness has many applications such as question …","url":["https://arxiv.org/pdf/2110.04845"]}
{"year":"2021","title":"What Transformers Might Know About the Physical World: T5 and the Origins of Knowledge","authors":["H Shi, P Wolff - Proceedings of the Annual Meeting of the Cognitive …, 2021"],"snippet":"… stack of decoders. All versions of T5 were trained on a cleaned version of the common crawl called the Colossal Cleaned Common Crawl (C4). The training corpus is over two times larger than Wikipedia. The largest version …","url":["https://escholarship.org/content/qt0kr3t179/qt0kr3t179.pdf"]}
{"year":"2021","title":"What Vision-Language ModelsSee'when they See Scenes","authors":["M Cafagna, K van Deemter, A Gatt - arXiv preprint arXiv:2109.07301, 2021"],"snippet":"Page 1. What Vision-Language Models 'See' when they See Scenes Michele Cafagna1 Kees van Deemter2 Albert Gatt1,2 1University of Malta, Institute of Linguistics and Language Technology 2Universiteit Utrecht, Information …","url":["https://arxiv.org/pdf/2109.07301"]}
{"year":"2021","title":"What's in the Box? An Analysis of Undesirable Content in the Common Crawl Corpus","authors":["A Sasha Luccioni, JD Viviano - arXiv e-prints, 2021","AS Luccioni, JD Viviano - arXiv preprint arXiv:2105.02732, 2021"],"snippet":"Whereas much of the success of the current generation of neural language models has been driven by increasingly large training corpora, relatively little research has been dedicated to analyzing these massive sources of textual data. In this …","url":["https://arxiv.org/pdf/2105.02732","https://ui.adsabs.harvard.edu/abs/2021arXiv210502732S/abstract"]}
{"year":"2021","title":"When FastText Pays Attention","authors":["V Novotný, M Štefánik, EF Ayetiran, P Sojka - 2021"],"snippet":"… (Hyper)-Parameter Optimization 3. Experiments & 4. Results I 3.3. (Hyper-)Parameter Optimization Dataset Number of tokens 2017 English Wikipedia 2,423,655,228 English Common Crawl 823,575,128,431 Table: Our datasets and their sizes in tokens …","url":["https://nlp.fi.muni.cz/trac/research/chrome/site/seminar2020/VNovotny-FastText_Attention_slides.pdf"]}
{"year":"2021","title":"When FastText Pays Attention: Efficient Estimation of Word Representations using Constrained Positional Weighting","authors":["V Novotný, M Štefánik, EF Ayetiran, P Sojka - arXiv preprint arXiv:2104.09691, 2021"],"snippet":"… For qualitative evaluation and performance estimation, we use the deduplicated English Common Crawl4 as the training corpus. See the statistics in Table 1. Dataset Number of tokens 2017 English Wikipedia 2,423,655,228 English Common Crawl 823,575,128,431 …","url":["https://arxiv.org/pdf/2104.09691"]}
{"year":"2021","title":"When few-shot learning meets large-scale knowledge-enhanced pre-training: Alibaba at FewCLUE","authors":["Z Xu, C Wang, P Li, Y Li, M Wang, B Hou, M Qiu… - … International Conference on …, 2021"],"snippet":"… In the pre-training steps, we use CLUECorpus2020 [28], which contains 100 GB Chinese raw texts retrieved from Common Crawl. We also employ 100 million high-quality Chinese knowledge SPOs from our in-house KG as the structured knowledge source …","url":["https://link.springer.com/chapter/10.1007/978-3-030-88483-3_34"]}
{"year":"2021","title":"When in Doubt, Summon the Titans: Efficient Inference with Large Models","authors":["AS Rawat, M Zaheer, AK Menon, A Ahmed, S Kumar - arXiv preprint arXiv …, 2021"],"snippet":"Scaling neural networks to \"large\" sizes, with billions of parameters, has been shown to yield impressive results on many challenging problems. However, the inference cost incurred by such large models often prevents their application in most …","url":["https://arxiv.org/pdf/2110.10305"]}
{"year":"2021","title":"When is BERT Multilingual? Isolating Crucial Ingredients for Cross-lingual Transfer","authors":["A Deshpande, P Talukdar, K Narasimhan - arXiv preprint arXiv:2110.14782, 2021"],"snippet":"… (same) which uses different splits of Wikipedia data (hence, same domain), and (2) non-parallel (diff) which uses Wikipedia data for the original and common crawl data (web text) for the derived language (hence, different domain). We use the …","url":["https://arxiv.org/pdf/2110.14782"]}
{"year":"2021","title":"Which Apple Keeps Which Doctor Away Colorful Word Representations with Visual Oracles","authors":["Z Zhang, H Yu, H Zhao, M Utiyama - IEEE/ACM Transactions on Audio, Speech, and …, 2021"],"snippet":"Recent pre-trained language models (PrLMs) offer a new performant method of contextualized word representations by leveraging the sequence-level context for modeling. Although the PrLMs generally provide more effective contextualized word …","url":["https://ieeexplore.ieee.org/abstract/document/9627795/"]}
{"year":"2021","title":"Which is Better for Deep Learning: Python or MATLAB? Answering Comparative Questions in Natural Language","authors":["V Chekalina, A Bondarenko, C Biemann, M Beloucif…"],"snippet":"… with respect to an aspect specified by the user. First, using the Elasticsearch BM25, CAM retrieves sentences containing the two compared objects and the comparison aspect from the Common Crawl-based corpus featuring …","url":["https://www.inf.uni-hamburg.de/en/inst/ab/lt/publications/2021-chekalinaetal-eacldemo-coqas.pdf"]}
{"year":"2021","title":"Which model is helpful in solving privacy, memorization, and bias problems?","authors":["S Yoon, H Lee"],"snippet":"Page 1. Which model is helpful in solving privacy, memorization, and bias problems? Soyoung Yoon * 1 Hyunji Lee * 1 Abstract Recently, language models of huge size are achieving stunning performance on natural language understanding tasks …","url":["https://soyoung97.github.io/profile/assets/papers/CS774.pdf"]}
{"year":"2021","title":"Who Blames or Endorses Whom? Entity-to-Entity Directed Sentiment Extraction in News Text","authors":["K Park, Z Pan, J Joo - arXiv preprint arXiv:2106.01033, 2021"],"snippet":"… For each of the target media outlets, we collected news articles shared throughout 2020 until September from the Common Crawl corpus, which has been collecting web data since 20088. The total number of retrieved …","url":["https://arxiv.org/pdf/2106.01033"]}
{"year":"2021","title":"Who is Accountable for Data Bias?","authors":["C Parkey - CHANCE, 2021"],"snippet":"… The OpenAI researchers use the data from Common Crawl, which provides a corpus for collaborative research, analysis, and education. It is an open data set containing petabytes of data collected since 2008 from raw web …","url":["https://www.tandfonline.com/doi/abs/10.1080/09332480.2021.1915032"]}
{"year":"2021","title":"Why Do Masked Neural Language Models Still Need Semantic Knowledge in Question Answering?","authors":["C Kang - 2021"],"snippet":"… Lastly, RoBERTa is pre-trained with much larger data. Especially, the data is about 160GB, which is composed of CommonCrawl News dataset [40], Open WebText corpus [41], and STORIES corpus [42] in addition to the pre-training data of BERT …","url":["http://sailab.kaist.ac.kr/wp-content/uploads/2021/03/MS_Thesis_cheongwoong.pdf"]}
{"year":"2021","title":"Wide and Deep Learning in Crowdfunding Success Prediction","authors":["D George"],"snippet":"The thesis studies to which extent deep learning models can predict whether crowdfunding campaigns will be successfully funded by drawing upon a dataset of 246,891 projects and 40 structured and text features from the crowdfunding platform …","url":["https://research.cbs.dk/files/68329943/1186110_Master_Thesis_133104_Wide_and_Deep_Learning_in_Crowdfunding_Success_Prediction.pdf"]}
{"year":"2021","title":"with Graph Neural Networks","authors":["M House, M Rules"],"snippet":"… example “pen” and “pencil”, or “salt” and “pepper”. We load word embeddings from a FastText 155 model [18], pre-trained on the Common Crawl dataset of over 600 billion tokens. Names of objects 156 are converted to word …","url":["https://openreview.net/pdf?id=Ei3MOY2rDHB"]}
{"year":"2021","title":"WITS: Wikipedia for Italian Text Summarization","authors":["S Casola, A Lavelli - 2021"],"snippet":"Abstractive text summarization has recently improved its performance due to the use of sequence to sequence models. However, while these models are extremely data-hungry, datasets in languages other than English are few. In this work, we introduce WITS (Wikipedia …","url":["http://ceur-ws.org/Vol-3033/paper65.pdf"]}
{"year":"2021","title":"Word embeddings for the classification of main conditions with samples of patient text describing symptoms","authors":["A Lamproudis - 2020"],"snippet":"Page 1. IN DEGREE PROJECT COMPUTER SCIENCE AND ENGINEERING, SECOND CYCLE, 30 CREDITS , STOCKHOLM SWEDEN 2020 Word embeddings for the classification of main conditions with samples of patient text describing symptoms …","url":["https://www.diva-portal.org/smash/get/diva2:1523304/FULLTEXT01.pdf"]}
{"year":"2021","title":"Word embeddings reveal social group attitudes and stereotypes in large language corpora","authors":["TES Charlesworth, MR Banaji"],"snippet":"… Surprisingly, less than half of the 45 languages studied by DeFranza and colleagues revealed a male-good/female-bad2 association across data from Wikipedia and the Common Crawl. This … For example, in the Common Crawl, 67% of languages with …","url":["http://www.people.fas.harvard.edu/~banaji/research/publications/articles/2021_Charlesworth_TAL.pdf"]}
{"year":"2021","title":"Word Vector Representations using Shallow Neural Networks","authors":["O Adewumi - 2021"],"snippet":"… We obtain better performance in both languages on the downstream task with far smaller training data, compared to recently released, common crawl versions; and character n-grams appear useful for Swedish, a morphologically rich language. Keywords …","url":["https://www.diva-portal.org/smash/record.jsf?pid=diva2:1543295"]}
{"year":"2021","title":"Word-level Korean-English Quality Estimation","authors":["S Eo, C Park, J Seo, H Moon, H Lim - Annual Conference on Human and Language …, 2021"],"snippet":"Abstract 기계번역 품질 예측 (Quality Estimation, QE) 은 정답 문장에 대한 참조없이 소스 문장과 기계번역 결과를 통해 기계번역 결과에 대한 품질을 수준별 주석으로 나타내주는 태스크이며, 다양한 활용도가 있다는 점에서 꾸준히 연구가 수행되고 있다 …","url":["https://www.koreascience.or.kr/article/CFKO202130060780846.pdf"]}
{"year":"2021","title":"Words as a window: Using word embeddings to explore the learned representations of Convolutional Neural Networks","authors":["D Dharmaretnam, C Foster, A Fyshe - Neural Networks, 2021"],"snippet":"… here. Elmo is trained on a random sample of Wikipedia and the common crawl, and we use an average of the three 512-dimensional output layers … 2016). The Lexvec model has 300 dimensions and is trained on the common crawl …","url":["https://www.sciencedirect.com/science/article/pii/S0893608020304263"]}
{"year":"2021","title":"WTR: A Test Collection for Web Table Retrieval","authors":["Z Chen, S Zhang, BD Davison - arXiv e-prints, 2021"],"snippet":"… PA, USA ABSTRACT We describe the development, characteristics and availability of a test collection for the task of Web table retrieval, which uses a largescale Web Table Corpora extracted from the Common Crawl. Since …","url":["https://arxiv.org/pdf/2105.02354"]}
{"year":"2021","title":"XLM-E: Cross-lingual Language Model Pre-training via ELECTRA","authors":["Z Chi, S Huang, L Dong, S Ma, S Singhal, P Bajaj… - arXiv preprint arXiv …, 2021"],"snippet":"… different languages. 4 Experiments 4.1 Setup Data We use the CC-100 (Conneau et al., 2020) dataset for the replaced token detection task. CC- 100 contains texts in 100 languages collected from the CommonCrawl dump. We use …","url":["https://arxiv.org/pdf/2106.16138"]}
{"year":"2021","title":"XLM-K: Improving Cross-Lingual Language Model Pre-Training with Multilingual Knowledge","authors":["X Jiang, Y Liang, W Chen, N Duan - arXiv preprint arXiv:2109.12573, 2021"],"snippet":"… 4.1 Implementation Details Data and Model Structure For the multilingual masked language modeling task, we use Common Crawl dataset (Wenzek et al. 2020) … 2020) was trained on Common Crawl dataset (Wenzek et al. 2020), which covers our training data Wikipedia …","url":["https://arxiv.org/pdf/2109.12573"]}
{"year":"2021","title":"XLM-T: A Multilingual Language Model Toolkit for Twitter","authors":["F Barbieri, LE Anke, J Camacho-Collados - arXiv preprint arXiv:2104.12250, 2021"],"snippet":"… 2.1). For the monolingual experiments we also include a FastText (FT) baseline (Joulin et al., 2017), which relies on monolingual FT embeddings trained on Common Crawl and Wikipedia (Grave et al., 2018) as initialization for each language lookup table …","url":["https://arxiv.org/pdf/2104.12250"]}
{"year":"2021","title":"XLM-T: Scaling up Multilingual Machine Translation with Pretrained Cross-lingual Transformer Encoders","authors":["S Ma, J Yang, H Huang, Z Chi, L Dong, D Zhang… - arXiv preprint arXiv …, 2020"],"snippet":"… cross-lingual representations. In this work, we adopt XLM-R BASE (Conneau et al., 2020) as the pretrained encoder. It was trained in 100 languages, using more than two terabytes of filtered CommonCrawl data. XLM-R is …","url":["https://arxiv.org/pdf/2012.15547"]}
{"year":"2021","title":"XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale","authors":["A Babu, C Wang, A Tjandra, K Lakhotia, Q Xu, N Goyal… - arXiv preprint arXiv …, 2021"],"snippet":"This paper presents XLS-R, a large-scale model for cross-lingual speech representation learning based on wav2vec 2.0. We train models with up to 2B parameters on nearly half a million hours of publicly available speech audio in 128 …","url":["https://arxiv.org/pdf/2111.09296"]}
{"year":"2021","title":"XRR: Explainable Risk Ranking for Financial Reports","authors":["TW Lin, RY Sun, HL Chang, CJ Wang, MF Tsai"],"snippet":"… permutation test with p < 0.05. Table 1: Performance comparison 3. GloVe [20] representations are 300-dimensional word vectors6 trained on 840 billion tokens of Common Crawl data. In the following experiments, we denote …","url":["https://2021.ecmlpkdd.org/wp-content/uploads/2021/07/sub_506.pdf"]}
{"year":"2021","title":"XTREME-R: Towards More Challenging and Nuanced Multilingual Evaluation","authors":["S Ruder, N Constant, J Botha, A Siddhant, O Firat, J Fu… - arXiv preprint arXiv …, 2021"],"snippet":"… Rebalanced mBERT (RemBERT; Chung et al., 2021) is a more efficient, scaled-up reparameterization of mBERT. These models have been pre-trained on unlabeled data in around 100 languages from Wikipedia …","url":["https://arxiv.org/pdf/2104.07412"]}
{"year":"2021","title":"XtremeDistilTransformers: Task Transfer for Task-agnostic Distillation","authors":["S Mukherjee, AH Awadallah, J Gao - arXiv preprint arXiv:2106.04563, 2021"],"snippet":"… To address these issues, we ex- plore techniques to automatically generate largescale task-specific transfer sets leveraging a very large bank of web sentences from Common Crawl in Section 4.1. 3 Distillation Framework Overview …","url":["https://arxiv.org/pdf/2106.04563"]}
{"year":"2021","title":"You too Brutus! Trapping Hateful Users in Social Media: Challenges, Solutions & Insights","authors":["M Das, P Saha, R Dutt, P Goyal, A Mukherjee… - arXiv preprint arXiv …, 2021"],"snippet":"… We experiment with the following text-classifiers. fastText: We leverage the pre-trained fastText embedding by Grave et al. [14] which has been trained on Common Crawl and Wikipedia. Each user's document has been represented as a 100 dimensional vector …","url":["https://arxiv.org/pdf/2108.00524"]}
{"year":"2021","title":"Yuan 1.0: Large-Scale Pre-trained Language Model in Zero-Shot and Few-Shot Learning","authors":["S Wu, X Zhao, T Yu, R Zhang, C Shen, H Liu, F Li… - arXiv preprint arXiv …, 2021"],"snippet":"… c) After this step, the data size of Common Crawl is decreased from 866,304GB to 12,200GB. • Traditional to Simplified Chinese: Converts traditional Chinese characters to simplified Chinese characters. • Sensitive Words Filtering: Removes …","url":["https://arxiv.org/pdf/2110.04725"]}
{"year":"2021","title":"Zero-Resource Multi-Dialectal Arabic Natural Language Understanding","authors":["M Khalifa, H Hassan, A Fahmy"],"snippet":"… XLM-R is a cross-lingual model, and we choose it since it is reported to perform better than mBERT, the multilingual model from Google [5]. XLM-R also uses Common Crawl for training, which is more likely to have …","url":["https://www.researchgate.net/profile/Muhammad-Khalifa/publication/350530331_Zero-Resource_Multi-Dialectal_Arabic_Natural_Language_Understanding/links/6064d0ba458515614d272082/Zero-Resource-Multi-Dialectal-Arabic-Natural-Language-Understanding.pdf"]}
{"year":"2021","title":"Zeroshot Crosslingual Transfer of a Gloss Language Model for Semantic Change Detection","authors":["M Rachinskiy, N Arefyev"],"snippet":"… finetuned. data with MLM (Masked Language Modeling) objective. But in contrast to BERT and RoBERTa, it is pretrained not on monolingual data but on 2.5 TB of texts from CommonCrawl in 100 languages. 3 System overview …","url":["http://www.dialog-21.ru/media/5278/rachinskiymplusarefyevn130.pdf"]}
{"year":"2021","title":"ZmBART: An Unsupervised Cross-lingual Transfer Framework for Language Generation","authors":["KK Maurya, MS Desarkar, Y Kano, K Deepshikha - arXiv preprint arXiv:2106.01597, 2021"],"snippet":"… languages. It is a transformer-based sequence- to-sequence pre-trained model. The model is trained on monolingual data in many languages from Wikipedia Common Crawl corpus with BART language model objective. Particularly …","url":["https://arxiv.org/pdf/2106.01597"]}