Datasets:
ilsp
/

Modalities:
Tabular
Text
Formats:
csv
Libraries:
Datasets
Dask
License:
File size: 2,791 Bytes
d08f312
 
e17c4e7
7eb6997
e17c4e7
7eb6997
c2ebf90
7eb6997
 
 
c2ebf90
7eb6997
 
 
c2ebf90
7eb6997
 
 
c2ebf90
7eb6997
 
 
c2ebf90
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d08f312
3698148
 
2634018
 
3698148
 
 
986c60a
3698148
986c60a
3698148
986c60a
3698148
 
986c60a
 
3698148
 
 
 
 
 
 
 
 
 
 
2634018
3698148
 
 
 
 
 
 
 
 
 
 
986c60a
3698148
 
 
 
 
 
986c60a
 
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
---
license: cc-by-nc-sa-4.0
configs:
- config_name: en-de
  data_files:
  - split: train
    path: en-de/*.tsv
- config_name: en-el
  data_files:
  - split: train
    path: en-el/*.tsv
- config_name: en-es
  data_files:
  - split: train
    path: en-es/*.tsv
- config_name: en-fr
  data_files:
  - split: train
    path: en-fr/*.tsv
- config_name: en-it
  data_files:
  - split: train
    path: en-it/*.tsv
task_categories:
- text-generation
- text2text-generation
- translation
- summarization
language:
- en
- el
- de
- fr
- es
- it
size_categories:
- 100K<n<1M
---
# SciPar Parallel Documents

<br/>

# Dataset Description

This dataset contains parallel documents (i.e., titles & abstracts) extracted from academic theses, dissertations, and other scientific texts.

In the original [paper](https://aclanthology.org/2022.lrec-1.284.pdf), we've extracted 9.17M sentence pairs in 31 language pairs from 86 repositories.

This version has been created through further processing and filtering to extract parallel **documents** instead of parallel **sentences**. 

To do this, we kept only the parallel titles and abstracts of academic records with high (mean) alignment scores by also using other filters (e.g., word ratio, non-empty abstracts, etc.).

Please note that this dataset will keep getting updated to include *more language pairs*, per the original paper.

<br/>

Table 1: Number of parallel documents per language pair

| Lang. Pair | # Parallel Docs |
|------------|-----------------|
| EN-DE      | 57,387          |
| EN-EL      | 55,833          |
| EN-ES      | 25,844          |
| EN-FR      | 130,750         |
| EN-IT      | 3,860           |
| **TOTAL**      | **273,674**         |

<br/>

# Related Datasets

The original datasets can be found in ELRC-SHARE:
- [Original SciPar in Moses format](https://elrc-share.eu/repository/browse/scipar-a-collection-of-parallel-corpora-from-scientific-abstracts-v-2021-in-moses-format/e78f32fe739611ec9c1a00155d026706a5d3e42019af467193538b102cf080f8/)
- [Original SciPar in TMX format](https://elrc-share.eu/repository/browse/scipar-a-collection-of-parallel-corpora-from-scientific-abstracts-v-2021-in-tmx-format/aaf503c0739411ec9c1a00155d02670665aacff53a8543938cd99da54fdd66af/)
- [SciPar UK-EN-RU in TMX format](https://elrc-share.eu/repository/browse/scipar-uk-en-ru/f635552ab06011ec9c1a00155d0267061ce92362f8af4c0b9d4f64d017c2df3f/)

# Citation
```
@inproceedings{roussis2022scipar,
  title={SciPar: A collection of parallel corpora from scientific abstracts},
  author={Roussis, Dimitrios and Papavassiliou, Vassilis and Prokopidis, Prokopis and Piperidis, Stelios and Katsouros, Vassilis},
  booktitle={Proceedings of the Thirteenth Language Resources and Evaluation Conference},
  pages={2652--2657},
  year={2022}
}
```