Datasets:

Modalities:
Text
Formats:
csv
ArXiv:
Libraries:
Datasets
pandas
License:
PMIndiaSum / README.md
pinzhenchen's picture
Update README.md
2416d5d
|
raw
history blame
No virus
5.59 kB
metadata
license: cc-by-4.0
task_categories:
  - summarization
language:
  - as
  - bn
  - gu
  - hi
  - mr
  - ml
  - mni
  - kn
  - pa
  - ta
  - or
  - te
  - ur
  - en
size_categories:
  - 100K<n<1M

Dataset Card for "PMIndiaSum"

Dataset Description

Summary

PMIndiaSum is a new multilingual and massively parallel headline summarization corpus focused on languages in India. Our corpus covers four language families, 14 languages, and the largest to date, 196 language pairs. It provides a testing ground for all cross-lingual pairs.

Supported tasks

Monolingual, multilingual and cross-lingual summarization for languages in India.

Languages

Assamese, Bengali, Gujarati, Hindi, Kannada, Marathi, Malayalam, Manipuri, Punjabi, Oriya, Telugu, Tamil, Urdu, English

Example Usage

Monolingual and cross-lingual summarization

Multilingual summarization

Dataset Structure

Data instances

We show an example of a Telugu-Hindi cross-lingual pair from PMIndiaSum:

{
"source_url": "https://www.pmindia.gov.in/te/news_updates/%E0%B0%8E%E0%B0%B2%E0%B0%95%E0%B1%8D%E0%B0%9F%E0%B1%8D%E0%B0%B0%E0%B0%BE%E0%B0%A8%E0%B0%BF%E0%B0%95%E0%B1%8D%E0%B0%B8%E0%B1%8D-%E0%B0%87%E0%B0%82%E0%B0%95%E0%B0%BE-%E0%B0%B8%E0%B0%AE%E0%B0%BE/"
"target_url": "https://www.pmindia.gov.in/hi/news_updates/%E0%A4%AA%E0%A5%8D%E0%A4%B0%E0%A4%A7%E0%A4%BE%E0%A4%A8%E0%A4%AE%E0%A4%82%E0%A4%A4%E0%A5%8D%E0%A4%B0%E0%A5%80-%E0%A4%B6%E0%A5%8D%E0%A4%B0%E0%A5%80-%E0%A4%A8%E0%A4%B0%E0%A5%87%E0%A4%A8%E0%A5%8D-45/"
"text": "ఎలక్ట్రానిక్స్, ఇంకా సమాచార సాంకేతిక విజ్ఞానం రంగంలో ద్వైపాక్షిక సహకారాన్ని పెంపొందింపచేయడంలో భారతదేశానికి మరియు అంగోలా కు మధ్య అవగాహనపూర్వక ఒప్పందాన్ని (ఎమ్ఒయు ను) గురించి ప్రధాన మంత్రి శ్రీ నరేంద్ర మోదీ అధ్యక్షతన జరిగిన కేంద్ర మంత్రివర్గ సమావేశం దృష్టి కి తీసుకువ‌చ్చారు. ఈ ఎమ్ఒయు ఇ-గవర్నెన్స్, సమాచార సాంకేతిక విజ్ఞాన సంబంధ విద్య కు అవసరమైన మానవ వనరుల వికాసం, సమాచార భద్రత, ఎలక్ట్రానిక్స్ హార్డ్ వేర్ తయారీ, ఐటి ఎంబెడెడ్ సాఫ్ట్ వేర్ ఇండస్ట్రీ, టెలిమెడిసిన్ తదితర రంగాలలో సన్నిహిత సహకారాన్ని పెంపొందింపచేయడానికి ఉద్దేశించినటువంటిది"
"summary": "मंत्रिमंडल को इलेक्‍ट्रॉनिक्‍स एवं संचना प्रौद्योगिकी के क्षेत्र में द्विपक्षीय सहयोग के लिए भारत और अंगोला के बीच समझौता ज्ञापन से अवगत कराया गया"

}

Data fields

  • 'source_url': A string representing the source article URL
  • 'target_url': A string representing the target article URL
  • 'text': A string containing the article text
  • 'summary': A string containing the article summary

Creation Details

Data source

The data source is PMIndia with license information at here.

We also extensively used materials from the PMIndia parallel corpus and its code.

Data construction details

You can find more details in our paper.

Other Information

License

Our materials are licensed under CC BY 4.0. We also request that you respect the policies from the source website.

Materials

Citation

Our work will be published at EMNLP Findings 2023. If you use our code or data, please kindly cite the following:

@inproceedings{urlana-etal-2023-pmindiasum,
    title={{PMIndiaSum}: Multilingual and Cross-lingual Headline Summarization for Languages in {India}}, 
    author={Urlana, Ashok and Chen, Pinzhen and Zhao, Zheng and Cohen, Shay B. and Shrivastava, Manish and Haddow, Barry},
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
    url ={https://arxiv.org/abs/2305.08828},
    year={2023}
}

Contributors

Ashok Urlana, Pinzhen Chen, Zheng Zhao, Shay B. Cohen, Manish Shrivastava, Barry Haddow

Contact

Ashok Urlana (ashokurlana@gmail.com)