File size: 2,183 Bytes
8a32c91
a8944de
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8a32c91
 
a8944de
 
 
 
 
456b39f
a8944de
6202e2f
a8944de
6202e2f
a8944de
6202e2f
95ae2e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
---
language:
- en
widget:
- text: "The Italian Space Agency’s Light Italian CubeSat for Imaging of Asteroids, or LICIACube, will fly by Dimorphos to capture images and video of the impact plume as it sprays up off the asteroid and maybe even spy the crater it could leave behind."
tags:
- seq2seq
- relation-extraction
- triple-generation
- entity-linking
- entity-type-linking
- relation-linking
model-index:
- name: knowgl
  results:
  - task:
      name: Relation Extraction
      type: Relation-Extraction
    dataset:
      name: "Babelscape/rebel-dataset"
      type: REBEL
    metrics:
       - name: RE+ Macro F1 
         type: re+ macro f1
         value: 70.74
license: cc-by-nc-sa-4.0
---

# KnowGL: Knowledge Generation and Linking from Text

The `knowgl-large` model is trained by combining Wikidata with an extended version of the training data [REBEL](https://huggingface.co/datasets/Babelscape/rebel-dataset) dataset. Given a sentence, it generates triple(s) in the following format -
```
[(subject mention # subject label # subject type) | relation label | (object mention # object label # object type)]
```
If there is more than one triple generated, they are separated by `$` in the output.

The model achieves state-of-the-art results for relation extraction on the REBEL dataset. See results in [Mihindukulasooriya et al (ISWC 2022)](https://arxiv.org/pdf/2207.05188.pdf).

The generated labels (for the subject, relation, and object) and their types can be directly mapped to Wikidata IDs associated with them.

#### Citation
```bibtex
@article{DBLP:journals/corr/abs-2207-05188,
  author    = {Nandana Mihindukulasooriya and
               Mike Sava and
               Gaetano Rossiello and
               Md. Faisal Mahbub Chowdhury and
               Irene Yachbes and
               Aditya Gidh and
               Jillian Duckwitz and
               Kovit Nisar and
               Michael Santos and
               Alfio Gliozzo},
  title     = {Knowledge Graph Induction enabling Recommending and Trend Analysis:
               {A} Corporate Research Community Use Case},
  journal   = {CoRR},
  volume    = {abs/2207.05188},
  year      = {2022}
}
```