annotations_creators:
- no-annotation
language_creators:
- expert-generated
- machine-generated
language:
- en
license:
- ms-pl
multilinguality:
- monolingual
- translation
size_categories:
- 1K<n<10K
source_datasets:
- extended|other-newstest2017
task_categories:
- translation
task_ids: []
paperswithcode_id: null
pretty_name: MsrZhenTranslationParity
dataset_info:
features:
- name: Reference-HT
dtype: string
- name: Reference-PE
dtype: string
- name: Combo-4
dtype: string
- name: Combo-5
dtype: string
- name: Combo-6
dtype: string
- name: Online-A-1710
dtype: string
splits:
- name: train
num_bytes: 1797033
num_examples: 2001
download_size: 0
dataset_size: 1797033
Dataset Card for msr_zhen_translation_parity
Table of Contents
- Dataset Description
- Dataset Structure
- Dataset Creation
- Considerations for Using the Data
- Additional Information
Dataset Description
- Homepage:
- Repository:
- Paper:
Achieving Human Parity on Automatic Chinese to English News Translation
- Leaderboard:
- Point of Contact:
Dataset Summary
Human evaluation results and translation output for the Translator Human Parity Data release, as described in https://blogs.microsoft.com/ai/machine-translation-news-test-set-human-parity/
The Translator Human Parity Data release contains all human evaluation results and translations related to our paper "Achieving Human Parity on Automatic Chinese to English News Translation", published on March 14, 2018. We have released this data to
- allow external validation of our claim of having achieved human parity
- to foster future research by releasing two additional human references for the Reference-WMT test set.
The dataset includes:
two new references for Chinese-English language pair of WMT17, one based on human translation from scratch (Reference-HT), the other based on human post-editing (Reference-PE);
human parity translations generated by our research systems Combo-4, Combo-5, and Combo-6, as well as translation output from online machine translation service Online-A-1710, collected on October 16, 2017;
The data package provided with the study also includes (but not parsed and provided as workable features of this dataset) all data points collected in human evaluation campaigns.
Supported Tasks and Leaderboards
[More Information Needed]
Languages
This dataset contains 6 extra English translations to Chinese-English language pair of WMT17.
Dataset Structure
Data Instances
[More Information Needed]
Data Fields
As mentioned in the summary, this dataset provides 6 extra English translations of Chinese-English language pair of WMT17.
Data fields are named exactly like the associated paper for easier cross-referenceing.
Reference-HT: human translation from scrach.Reference-PE: human post-editing.Combo-4,Combo-5,Combo-6: three translations by research systems.Online-A-1710: a translation from an anonymous online machine translation service.
All data fields of a record are translations for the same Chinese source sentence.
Data Splits
[More Information Needed]
Dataset Creation
Curation Rationale
[More Information Needed]
Source Data
Initial Data Collection and Normalization
[More Information Needed]
Who are the source language producers?
[More Information Needed]
Annotations
Annotation process
[More Information Needed]
Who are the annotators?
[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
[More Information Needed]
Licensing Information
[More Information Needed]
Citation Information
Citation information is available at this link Achieving Human Parity on Automatic Chinese to English News Translation
Contributions
Thanks to @leoxzhao for adding this dataset.