Hugging Face's logo

Datasets:
gsarti
/
wmt_vat

Tasks:
Text Generation
Translation
Languages: Czech German English
Multilinguality: multilingual translation
Size Categories: unknown
Language Creators: expert-generated
Annotations Creators: found
Source Datasets: extended|wmt16 extended|wmt17 extended|wmt18
Tags: conditional-text-generation
License:
Dataset card Files Community
3
wmt_vat / README.md
gsarti's picture
gsarti
Fix task tags (#3)
986c40d
preview code
|
raw history blame contribute delete
No virus
11.3 kB
metadata
annotations_creators:
  - found
language_creators:
  - expert-generated
language:
  - cs
  - de
  - en
  - et
  - fi
  - fr
  - gu
  - iu
  - ja
  - kk
  - km
  - lt
  - lv
  - pl
  - ps
  - ro
  - ru
  - ta
  - tr
  - zh
license:
  - unknown
multilinguality:
  - multilingual
  - translation
size_categories:
  - unknown
source_datasets:
  - extended|wmt16
  - extended|wmt17
  - extended|wmt18
  - extended|wmt19
  - extended|wmt20
task_categories:
  - text-generation
  - translation
task_ids: []
pretty_name: wmt_vat
tags:
  - conditional-text-generation

Dataset Card for Variance-Aware MT Test Sets

Table of Contents

Dataset Description

Dataset Summary

This dataset comprises 70 small and discriminative test sets for machine translation (MT) evaluation called variance-aware test sets (VAT), covering 35 translation directions from WMT16 to WMT20 competitions. VAT is automatically created by a novel variance-aware filtering method that filters the indiscriminative test instances of the current MT benchmark without any human labor. Experimental results show that VAT outperforms the original WMT benchmark in terms of the correlation with human judgment across mainstream language pairs and test sets. Further analysis on the properties of VAT reveals the challenging linguistic features (e.g., translation of low-frequency words and proper nouns) for the competitive MT systems, providing guidance for constructing future MT test sets.

Disclaimer: *The VAT test sets are hosted through Github by the Natural Language Processing & Portuguese-Chinese Machine Translation Laboratory (NLP2CT Lab) of the University of Macau. They were introduced by the paper Variance-Aware Machine Translation Test Sets by Runzhe Zhan, Xuebo Liu, Derek F. Wong, Lidia S. Chao and follow the original licensing for WMT test sets.

Supported Tasks and Leaderboards

Machine Translation

Refer to the original paper for additional details on model evaluation on VAT.

Languages

The following table taken from the original paper lists the languages supported by the VAT test sets, for a total of 70 language pairs:

↔️ wmt16 wmt17 wmt18 wmt19 wmt20
xx_en cs,de,fi,
ro,ru,tr		 cs,de,fi,lv,
ru,tr,zh		 cs,de,et,fi,
ru,tr,zh		 de,fi,gu,
kk,lt,ru,zh		 cs,de,iu,ja,km,
pl,ps,ru,ta,zh
en_xx ru cs,de,fi,
lv,ru,tr,zh		 cs,de,et,fi,
ru,tr,zh		 cs,de,fi,gu,
kk,lt,ru,zh		 cs,de,ja,pl,
ru,ta,zh
xx_yy / / / de_cs,de_fr,
fr_de		 /

To use any one of the test set, pass wmtXX_src_tgt as configuration name to the load_dataset command. E.g. to load the English-Russian test set from wmt16, use load_dataset('gsarti/wmt_vat', 'wmt16_en_ru').

Dataset Structure

Data Instances

A sample from the test split (the only available split) for the WMT16 English-Russian language (wmt16_en_ru config) is provided below. All configurations have the same structure.

{
    'orig_id': 0,
    'source': 'The social card of residents of Ivanovo region is to be recognised as an electronic payment instrument.',
    'reference': 'Социальная карта жителя Ивановской области признается электронным средством платежа.'
}

The text is provided as-in the original dataset, without further preprocessing or tokenization.

Data Fields

  • orig_id: Id corresponding to the row id in the original dataset, before variance-aware filtering.
  • source: The source sentence.
  • reference: The reference sentence in the target language.

Data Splits

Taken from the original repository:

Configuration # Sentences # Words # Vocabulary
wmt20_km_en 928 17170 3645
wmt20_cs_en 266 12568 3502
wmt20_en_de 567 21336 5945
wmt20_ja_en 397 10526 3063
wmt20_ps_en 1088 20296 4303
wmt20_en_zh 567 18224 5019
wmt20_en_ta 400 7809 4028
wmt20_de_en 314 16083 4046
wmt20_zh_en 800 35132 6457
wmt20_en_ja 400 12718 2969
wmt20_en_cs 567 16579 6391
wmt20_en_pl 400 8423 3834
wmt20_en_ru 801 17446 6877
wmt20_pl_en 400 7394 2399
wmt20_iu_en 1188 23494 3876
wmt20_ru_en 396 6966 2330
wmt20_ta_en 399 7427 2148
wmt19_zh_en 800 36739 6168
wmt19_en_cs 799 15433 6111
wmt19_de_en 800 15219 4222
wmt19_en_gu 399 8494 3548
wmt19_fr_de 680 12616 3698
wmt19_en_zh 799 20230 5547
wmt19_fi_en 798 13759 3555
wmt19_en_fi 799 13303 6149
wmt19_kk_en 400 9283 2584
wmt19_de_cs 799 15080 6166
wmt19_lt_en 400 10474 2874
wmt19_en_lt 399 7251 3364
wmt19_ru_en 800 14693 3817
wmt19_en_kk 399 6411 3252
wmt19_en_ru 799 16393 6125
wmt19_gu_en 406 8061 2434
wmt19_de_fr 680 16181 3517
wmt19_en_de 799 18946 5340
wmt18_en_cs 1193 19552 7926
wmt18_cs_en 1193 23439 5453
wmt18_en_fi 1200 16239 7696
wmt18_en_tr 1200 19621 8613
wmt18_en_et 800 13034 6001
wmt18_ru_en 1200 26747 6045
wmt18_et_en 800 20045 5045
wmt18_tr_en 1200 25689 5955
wmt18_fi_en 1200 24912 5834
wmt18_zh_en 1592 42983 7985
wmt18_en_zh 1592 34796 8579
wmt18_en_ru 1200 22830 8679
wmt18_de_en 1199 28275 6487
wmt18_en_de 1199 25473 7130
wmt17_en_lv 800 14453 6161
wmt17_zh_en 800 20590 5149
wmt17_en_tr 1203 17612 7714
wmt17_lv_en 800 18653 4747
wmt17_en_de 1202 22055 6463
wmt17_ru_en 1200 24807 5790
wmt17_en_fi 1201 17284 7763
wmt17_tr_en 1203 23037 5387
wmt17_en_zh 800 18001 5629
wmt17_en_ru 1200 22251 8761
wmt17_fi_en 1201 23791 5300
wmt17_en_cs 1202 21278 8256
wmt17_de_en 1202 23838 5487
wmt17_cs_en 1202 22707 5310
wmt16_tr_en 1200 19225 4823
wmt16_ru_en 1199 23010 5442
wmt16_ro_en 800 16200 3968
wmt16_de_en 1200 22612 5511
wmt16_en_ru 1199 20233 7872
wmt16_fi_en 1200 20744 5176
wmt16_cs_en 1200 23235 5324

Dataset Creation

The dataset was created by retaining a subset of the top 40% instances from various WMT test sets for which the variance between automatic scores (BLEU, BLEURT, COMET, BERTScore) was the highest. Please refer to the original article Variance-Aware Machine Translation Test Sets for additional information on dataset creation.

Additional Information

Dataset Curators

The original authors of VAT are the curators of the original dataset. For problems or updates on this 🤗 Datasets version, please contact gabriele.sarti996@gmail.com.

Licensing Information

The variance-aware test set were created based on the original WMT test set. Thus, the the original data licensing plan already stated by WMT organizers is still applicable:

The data released for the WMT news translation task can be freely used for research purposes, we just ask that you cite the WMT shared task overview paper, and respect any additional citation requirements on the individual data sets. For other uses of the data, you should consult with original owners of the data sets.

Citation Information

Please cite the authors if you use these corpora in your work. It is also advised to cite the original WMT shared task paper for the specific test sets that were used.

@inproceedings{
    zhan2021varianceaware,
    title={Variance-Aware Machine Translation Test Sets},
    author={Runzhe Zhan and Xuebo Liu and Derek F. Wong and Lidia S. Chao},
    booktitle={Thirty-fifth Conference on Neural Information Processing Systems, Datasets and Benchmarks Track},
    year={2021},
    url={https://openreview.net/forum?id=hhKA5k0oVy5}
}