Datasets:
Tasks:
Text2Text Generation
Sub-tasks:
text-simplification
Languages:
English
Size:
100K<n<1M
ArXiv:
License:
File size: 15,387 Bytes
94f3afd b641cc6 94f3afd fb726e8 94f3afd fb726e8 b641cc6 94f3afd cb733c5 94f3afd 49fb980 0c980ff b641cc6 07ecdd5 97cacd6 07ecdd5 94f3afd 49fb980 94f3afd 49fb980 94f3afd da0bdb7 94f3afd b266476 94f3afd 7d55b4e 94f3afd 7d55b4e 94f3afd c2a371f 94f3afd 7d55b4e 94f3afd 7d55b4e 94f3afd 9c1e182 94f3afd da0bdb7 07ecdd5 |
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 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 |
---
annotations_creators:
- crowdsourced
- machine-generated
language_creators:
- found
language:
- en
license:
- cc-by-sa-3.0
multilinguality:
- monolingual
size_categories:
- 100K<n<1M
source_datasets:
- extended|other-wikipedia
task_categories:
- text2text-generation
task_ids:
- text-simplification
paperswithcode_id: null
pretty_name: WikiAuto
configs:
- auto
- auto_acl
- auto_full_no_split
- auto_full_with_split
- manual
dataset_info:
- config_name: manual
features:
- name: alignment_label
dtype:
class_label:
names:
0: notAligned
1: aligned
2: partialAligned
- name: normal_sentence_id
dtype: string
- name: simple_sentence_id
dtype: string
- name: normal_sentence
dtype: string
- name: simple_sentence
dtype: string
- name: gleu_score
dtype: float32
splits:
- name: train
num_bytes: 110838475
num_examples: 373801
- name: dev
num_bytes: 21112775
num_examples: 73249
- name: test
num_bytes: 33851634
num_examples: 118074
download_size: 168957430
dataset_size: 165802884
- config_name: auto_acl
features:
- name: normal_sentence
dtype: string
- name: simple_sentence
dtype: string
splits:
- name: full
num_bytes: 121975414
num_examples: 488332
download_size: 118068366
dataset_size: 121975414
- config_name: auto
features:
- name: example_id
dtype: string
- name: normal
struct:
- name: normal_article_id
dtype: int32
- name: normal_article_title
dtype: string
- name: normal_article_url
dtype: string
- name: normal_article_content
sequence:
- name: normal_sentence_id
dtype: string
- name: normal_sentence
dtype: string
- name: simple
struct:
- name: simple_article_id
dtype: int32
- name: simple_article_title
dtype: string
- name: simple_article_url
dtype: string
- name: simple_article_content
sequence:
- name: simple_sentence_id
dtype: string
- name: simple_sentence
dtype: string
- name: paragraph_alignment
sequence:
- name: normal_paragraph_id
dtype: string
- name: simple_paragraph_id
dtype: string
- name: sentence_alignment
sequence:
- name: normal_sentence_id
dtype: string
- name: simple_sentence_id
dtype: string
splits:
- name: part_1
num_bytes: 1773240295
num_examples: 125059
- name: part_2
num_bytes: 80417651
num_examples: 13036
download_size: 2160638921
dataset_size: 1853657946
- config_name: auto_full_no_split
features:
- name: normal_sentence
dtype: string
- name: simple_sentence
dtype: string
splits:
- name: full
num_bytes: 146310611
num_examples: 591994
download_size: 141574179
dataset_size: 146310611
- config_name: auto_full_with_split
features:
- name: normal_sentence
dtype: string
- name: simple_sentence
dtype: string
splits:
- name: full
num_bytes: 124549115
num_examples: 483801
download_size: 120678315
dataset_size: 124549115
---
# Dataset Card for WikiAuto
## Table of Contents
- [Dataset Description](#dataset-description)
- [Dataset Summary](#dataset-summary)
- [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
- [Languages](#languages)
- [Dataset Structure](#dataset-structure)
- [Data Instances](#data-instances)
- [Data Fields](#data-fields)
- [Data Splits](#data-splits)
- [Dataset Creation](#dataset-creation)
- [Curation Rationale](#curation-rationale)
- [Source Data](#source-data)
- [Annotations](#annotations)
- [Personal and Sensitive Information](#personal-and-sensitive-information)
- [Considerations for Using the Data](#considerations-for-using-the-data)
- [Social Impact of Dataset](#social-impact-of-dataset)
- [Discussion of Biases](#discussion-of-biases)
- [Other Known Limitations](#other-known-limitations)
- [Additional Information](#additional-information)
- [Dataset Curators](#dataset-curators)
- [Licensing Information](#licensing-information)
- [Citation Information](#citation-information)
- [Contributions](#contributions)
## Dataset Description
- **Repository:** [WikiAuto github repository](https://github.com/chaojiang06/wiki-auto)
- **Paper:** [Neural CRF Model for Sentence Alignment in Text Simplification](https://arxiv.org/abs/2005.02324)
- **Point of Contact:** [Chao Jiang](jiang.1530@osu.edu)
### Dataset Summary
WikiAuto provides a set of aligned sentences from English Wikipedia and Simple English Wikipedia as a resource to train sentence simplification systems.
The authors first crowd-sourced a set of manual alignments between sentences in a subset of the Simple English Wikipedia and their corresponding versions in English Wikipedia (this corresponds to the `manual` config in this version of dataset), then trained a neural CRF system to predict these alignments.
The trained alignment prediction model was then applied to the other articles in Simple English Wikipedia with an English counterpart to create a larger corpus of aligned sentences (corresponding to the `auto`, `auto_acl`, `auto_full_no_split`, and `auto_full_with_split` configs here).
### Supported Tasks and Leaderboards
The dataset was created to support a `text-simplification` task. Success in these tasks is typically measured using the [SARI](https://huggingface.co/metrics/sari) and [FKBLEU](https://huggingface.co/metrics/fkbleu) metrics described in the paper [Optimizing Statistical Machine Translation for Text Simplification](https://www.aclweb.org/anthology/Q16-1029.pdf).
### Languages
While both the input and output of the proposed task are in English (`en`), it should be noted that it is presented as a translation task where Wikipedia Simple English is treated as its own idiom. For a statement of what is intended (but not always observed) to constitute Simple English on this platform, see [Simple English in Wikipedia](https://simple.wikipedia.org/wiki/Wikipedia:About#Simple_English).
## Dataset Structure
### Data Instances
The data in all of the configurations looks a little different.
A `manual` config instance consists of a sentence from the Simple English Wikipedia article, one from the linked English Wikipedia article, IDs for each of them, and a label indicating whether they are aligned. Sentences on either side can be repeated so that the aligned sentences are in the same instances. For example:
```
{'alignment_label': 1,
'normal_sentence_id': '0_66252-1-0-0',
'simple_sentence_id': '0_66252-0-0-0',
'normal_sentence': 'The Local Government Act 1985 is an Act of Parliament in the United Kingdom.', 'simple_sentence': 'The Local Government Act 1985 was an Act of Parliament in the United Kingdom', 'gleu_score': 0.800000011920929}
```
Is followed by
```
{'alignment_label': 0,
'normal_sentence_id': '0_66252-1-0-1',
'simple_sentence_id': '0_66252-0-0-0',
'normal_sentence': 'Its main effect was to abolish the six county councils of the metropolitan counties that had been set up in 1974, 11 years earlier, by the Local Government Act 1972, along with the Greater London Council that had been established in 1965.',
'simple_sentence': 'The Local Government Act 1985 was an Act of Parliament in the United Kingdom', 'gleu_score': 0.08641975373029709}
```
The `auto` config shows a pair of an English and corresponding Simple English Wikipedia as an instance, with an alignment at the paragraph and sentence level:
```
{'example_id': '0',
'normal': {'normal_article_content': {'normal_sentence': ["Lata Mondal ( ; born: 16 January 1993, Dhaka) is a Bangladeshi cricketer who plays for the Bangladesh national women's cricket team.",
'She is a right handed batter.',
'Mondal was born on January 16, 1993 in Dhaka, Bangladesh.',
"Mondal made her ODI career against the Ireland women's cricket team on November 26, 2011.",
"Mondal made her T20I career against the Ireland women's cricket team on August 28, 2012.",
"In October 2018, she was named in Bangladesh's squad for the 2018 ICC Women's World Twenty20 tournament in the West Indies.",
"Mondal was a member of the team that won a silver medal in cricket against the China national women's cricket team at the 2010 Asian Games in Guangzhou, China."],
'normal_sentence_id': ['normal-41918715-0-0',
'normal-41918715-0-1',
'normal-41918715-1-0',
'normal-41918715-2-0',
'normal-41918715-3-0',
'normal-41918715-3-1',
'normal-41918715-4-0']},
'normal_article_id': 41918715,
'normal_article_title': 'Lata Mondal',
'normal_article_url': 'https://en.wikipedia.org/wiki?curid=41918715'},
'paragraph_alignment': {'normal_paragraph_id': ['normal-41918715-0'],
'simple_paragraph_id': ['simple-702227-0']},
'sentence_alignment': {'normal_sentence_id': ['normal-41918715-0-0',
'normal-41918715-0-1'],
'simple_sentence_id': ['simple-702227-0-0', 'simple-702227-0-1']},
'simple': {'simple_article_content': {'simple_sentence': ["Lata Mondal (born: 16 January 1993) is a Bangladeshi cricketer who plays for the Bangladesh national women's cricket team.",
'She is a right handed bat.'],
'simple_sentence_id': ['simple-702227-0-0', 'simple-702227-0-1']},
'simple_article_id': 702227,
'simple_article_title': 'Lata Mondal',
'simple_article_url': 'https://simple.wikipedia.org/wiki?curid=702227'}}
```
Finally, the `auto_acl`, the `auto_full_no_split`, and the `auto_full_with_split` configs were obtained by selecting the aligned pairs of sentences from `auto` to provide a ready-to-go aligned dataset to train a sequence-to-sequence system. While `auto_acl` corresponds to the filtered version of the data used to train the systems in the paper, `auto_full_no_split` and `auto_full_with_split` correspond to the unfiltered versions with and without sentence splits respectively. In the `auto_full_with_split` config, we join the sentences in the simple article mapped to the same sentence in the complex article to capture sentence splitting. Split sentences are separated by a `<SEP>` token. In the `auto_full_no_split` config, we do not join the splits and treat them as separate pairs. An instance is a single pair of sentences:
```
{'normal_sentence': 'In early work , Rutherford discovered the concept of radioactive half-life , the radioactive element radon , and differentiated and named alpha and beta radiation .\n',
'simple_sentence': 'Rutherford discovered the radioactive half-life , and the three parts of radiation which he named Alpha , Beta , and Gamma .\n'}
```
### Data Fields
The data has the following field:
- `normal_sentence`: a sentence from English Wikipedia.
- `normal_sentence_id`: a unique ID for each English Wikipedia sentence. The last two dash-separated numbers correspond to the paragraph number in the article and the sentence number in the paragraph.
- `simple_sentence`: a sentence from Simple English Wikipedia.
- `simple_sentence_id`: a unique ID for each Simple English Wikipedia sentence. The last two dash-separated numbers correspond to the paragraph number in the article and the sentence number in the paragraph.
- `alignment_label`: signifies whether a pair of sentences is aligned: labels are `2:partialAligned`, `1:aligned` and `0:notAligned`
- `paragraph_alignment`: a first step of alignment mapping English and Simple English paragraphs from linked articles
- `sentence_alignment`: the full alignment mapping English and Simple English sentences from linked articles
- `gleu_score`: the sentence level GLEU (Google-BLEU) score for each pair.
### Data Splits
In `auto`, the `part_2` split corresponds to the articles used in `manual`, and `part_1` has the rest of Wikipedia.
The `manual` config is provided with a `train`/`dev`/`test` split with the following amounts of data:
| | train | validation | test |
|------------------------|--------:|-----------:|--------:|
| Total sentence pairs | 373801 | 73249 | 118074 |
| Aligned sentence pairs | 1889 | 346 | 677 |
## Dataset Creation
### Curation Rationale
Simple English Wikipedia provides a ready source of training data for text simplification systems, as 1. articles in different languages are linked, making it easier to find parallel data and 2. the Simple English data is written by users for users rather than by professional translators. However, even though articles are aligned, finding a good sentence-level alignment can remain challenging. This work aims to provide a solution for this problem. By manually annotating a sub-set of the articles, they manage to achieve an F1 score of over 88% on predicting alignment, which allows to create a good quality sentence level aligned corpus using all of Simple English Wikipedia.
### Source Data
#### Initial Data Collection and Normalization
The authors mention that they "extracted 138,095 article pairs from the 2019/09 Wikipedia dump [...] using an improved version of the [WikiExtractor](https://github.com/attardi/wikiextractor) library". The [SpaCy](https://spacy.io/) library is used for sentence splitting.
#### Who are the source language producers?
The dataset uses langauge from Wikipedia: some demographic information is provided [here](https://en.wikipedia.org/wiki/Wikipedia:Who_writes_Wikipedia%3F).
### Annotations
#### Annotation process
Sentence alignment labels were obtained for 500 randomly sampled document pairs (10,123 sentence pairs total). The authors pre-selected several alignment candidates from English Wikipedia for each Simple Wikipedia sentence based on various similarity metrics, then asked the crowd-workers to annotate these pairs.
#### Who are the annotators?
No demographic annotation is provided for the crowd workers.
[More Information Needed]
### Personal and Sensitive Information
[More Information Needed]
## Considerations for Using the Data
### Social Impact of Dataset
[More Information Needed]
### Discussion of Biases
[More Information Needed]
### Other Known Limitations
[More Information Needed]
## Additional Information
### Dataset Curators
The dataset was created by Chao Jiang, Mounica Maddela, Wuwei Lan, Yang Zhong, and Wei Xu working at Ohio State University.
### Licensing Information
The dataset is not licensed by itself, but the source Wikipedia data is under a `cc-by-sa-3.0` license.
### Citation Information
You can cite the paper presenting the dataset as:
```
@inproceedings{acl/JiangMLZX20,
author = {Chao Jiang and
Mounica Maddela and
Wuwei Lan and
Yang Zhong and
Wei Xu},
editor = {Dan Jurafsky and
Joyce Chai and
Natalie Schluter and
Joel R. Tetreault},
title = {Neural {CRF} Model for Sentence Alignment in Text Simplification},
booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational
Linguistics, {ACL} 2020, Online, July 5-10, 2020},
pages = {7943--7960},
publisher = {Association for Computational Linguistics},
year = {2020},
url = {https://www.aclweb.org/anthology/2020.acl-main.709/}
}
```
### Contributions
Thanks to [@yjernite](https://github.com/yjernite), [@mounicam](https://github.com/mounicam) for adding this dataset. |