de-francophones
commited on
Commit
•
667340e
1
Parent(s):
21264c3
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
- fr
|
5 |
+
- de
|
6 |
+
- es
|
7 |
+
- tr
|
8 |
+
configs:
|
9 |
+
- config_name: en
|
10 |
+
data_files:
|
11 |
+
- split: train
|
12 |
+
path: "RELX_en.json"
|
13 |
+
- config_name: fr
|
14 |
+
data_files:
|
15 |
+
- split: train
|
16 |
+
path: "RELX_fr.json"
|
17 |
+
- config_name: de
|
18 |
+
data_files:
|
19 |
+
- split: train
|
20 |
+
path: "RELX_de.json"
|
21 |
+
- config_name: es
|
22 |
+
data_files:
|
23 |
+
- split: train
|
24 |
+
path: "RELX_es.json"
|
25 |
+
- config_name: tr
|
26 |
+
data_files:
|
27 |
+
- split: train
|
28 |
+
path: "RELX_tr.json"
|
29 |
+
---
|
30 |
+
|
31 |
+
> [!NOTE]
|
32 |
+
> Dataset origin: https://github.com/boun-tabi/RELX
|
33 |
+
|
34 |
+
|
35 |
+
## RELX-Distant
|
36 |
+
|
37 |
+
This dataset is gathered from Wikipedia and Wikidata.
|
38 |
+
The process is as follows:
|
39 |
+
|
40 |
+
1. The Wikipedia dumps for the corresponding languages are downloaded and converted into raw documents with Wikipedia hyperlinks in entities.
|
41 |
+
2. The raw documents are split into sentences with spaCy (Honnibal and Montani, 2017), and all hyperlinks are converted to their corresponding Wikidata IDs.
|
42 |
+
3. Sentences that include entity pairs with Wikidata relations (Vrandečić and Krötzsch, 2014) are collected. We filter and combine some of the relations and propose RELX-Distant whose statistics can be seen in the table below.
|
43 |
+
|
44 |
+
| **Language** | **Number of Sentences** |
|
45 |
+
|--------------|-------------------------|
|
46 |
+
| English | 815689 |
|
47 |
+
| French | 652842 |
|
48 |
+
| German | 652062 |
|
49 |
+
| Spanish | 397875 |
|
50 |
+
| Turkish | 57114 |
|
51 |
+
|
52 |
+
## Citation
|
53 |
+
|
54 |
+
```
|
55 |
+
@inproceedings{koksal-ozgur-2020-relx,
|
56 |
+
title = "The {RELX} Dataset and Matching the Multilingual Blanks for Cross-Lingual Relation Classification",
|
57 |
+
author = {K{\"o}ksal, Abdullatif and
|
58 |
+
{\"O}zg{\"u}r, Arzucan},
|
59 |
+
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
|
60 |
+
month = nov,
|
61 |
+
year = "2020",
|
62 |
+
address = "Online",
|
63 |
+
publisher = "Association for Computational Linguistics",
|
64 |
+
url = "https://www.aclweb.org/anthology/2020.findings-emnlp.32",
|
65 |
+
doi = "10.18653/v1/2020.findings-emnlp.32",
|
66 |
+
pages = "340--350",
|
67 |
+
}
|
68 |
+
```
|