FradSer/OpenSubtitles-en-zh-cn-20m

Dataset Card for OpenSubtitles English-Chinese 20M Parallel Corpus

This dataset contains English-Chinese parallel subtitles from OpenSubtitles, specifically the English to Simplified Chinese (zh-CN) language pair. The data is sourced from the OpenSubtitles v2024 collection available at OPUS.

Dataset Details

Dataset Description

This is a collection of parallel movie subtitles in English and Chinese, extracted from OpenSubtitles.org. The dataset features improved sentence alignment and enhanced language checking compared to previous versions. It includes various types of subtitle alignments:

• Standard parallel sentences between English and Chinese
• Alternative subtitle alignments with different variations
• Intra-lingual alignments showing different translations or subtitle versions

The dataset is particularly useful for:

• Machine translation research and development

• Cross-lingual studies

• Subtitle translation systems

• Language learning applications

• Curated by: OPUS - The Open Parallel Corpus

• Language(s) (NLP): English (en), Simplified Chinese (zh-CN)

• License: MIT

Dataset Sources

• Repository: https://opus.nlpl.eu/OpenSubtitles/en&zh-CN/v2024/OpenSubtitles
• Original Data Source: http://www.opensubtitles.org/
• Paper: P. Lison and J. Tiedemann, 2016, OpenSubtitles2016: Extracting Large Parallel Corpora from Movie and TV Subtitles. In Proceedings of the 10th International Conference on Language Resources and Evaluation (LREC 2016)

Uses

Direct Use

The dataset is suitable for:

• Training machine translation models
• Studying colloquial language patterns
• Developing subtitle translation systems
• Cross-lingual information retrieval
• Language learning applications

Out-of-Scope Use

The dataset should not be used for:

• Commercial redistribution without proper attribution
• Applications requiring formal or literary language translation
• Legal or medical translation purposes where professional translation is required

Dataset Structure

The dataset contains parallel sentences in English and Chinese, with various types of alignments:

• Direct translations
• Alternative translations
• Paraphrases
• Punctuation variations
• Spelling variations

Dataset Creation

Curation Rationale

The dataset was created to provide a large-scale parallel corpus for English-Chinese translation tasks, specifically focused on movie and TV show subtitles. The curation process emphasizes improved sentence alignment and language quality checking.

Source Data

Data Collection and Processing

The data was collected from OpenSubtitles.org and processed using:

• Improved sentence alignment techniques
• Language quality checking
• Alternative subtitle alignment processing
• Intra-lingual alignment categorization

Who are the source data producers?

The source data comes from subtitle creators and translators who contributed to OpenSubtitles.org. These include both professional and community translators.

Personal and Sensitive Information

The dataset contains only movie and TV show subtitles and does not include personal information of subtitle creators or translators.

Citation

BibTeX:

@InProceedings{LISON16.947,
  author = {Pierre Lison and Jörg Tiedemann},
  title = {OpenSubtitles2016: Extracting Large Parallel Corpora from Movie and TV Subtitles},
  booktitle = {Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC 2016)},
  year = {2016},
  publisher = {European Language Resources Association (ELRA)},
  address = {Portorož, Slovenia}
}

Dataset Card Contact

For more information about this dataset, please visit:

• OPUS Project: https://opus.nlpl.eu/
• OpenSubtitles: http://www.opensubtitles.org/