Mwnthai/bodo-legal-judgment-summarization

Bodo Legal Judgment Summarization Dataset

Overview

This dataset contains Bodo-language legal judgments paired with their corresponding summaries. It is intended for research on automatic legal text summarization, low-resource NLP, and Bodo language modeling.

Dataset Structure

Each example contains:

• —id: Document identifier corresponding to the original TXT filenames.
• —language: Language of the document (Bodo).
• —domain: Domain (Legal).
• —judgment: Full legal judgment text in Bodo.
• —summary: Corresponding summary in Bodo.

Data Provenance and Credit

This Bodo dataset is a translated/adapted version of the legal case document summarization data released by Shukla et al. (2022).

The original dataset is associated with the work:

Abhay Shukla, Paheli Bhattacharya, Soham Poddar, Rajdeep Mukherjee, Kripabandhu Ghosh, Pawan Goyal, and Saptarshi Ghosh. (2022). Legal Case Document Summarization: Extractive and Abstractive Methods and their Evaluation. Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (AACL-IJCNLP 2022), pp. 1048–1064. DOI: 10.18653/v1/2022.aacl-main.77

The original legal summarization dataset is available through Zenodo:

https://zenodo.org/records/7152317

The corresponding research paper is available through the ACL Anthology:

https://aclanthology.org/2022.aacl-main.77/

The original Zenodo repository describes three legal summarization datasets, including IN-Abs, which contains Indian Supreme Court case documents and abstractive summaries. This repository uses the relevant source material as the basis for a Bodo-language translation/adaptation.

Transformation

The original source material was translated into Bodo for research on low-resource Indian-language legal summarization. The Bodo text in this repository should therefore be considered a derived/translated dataset, rather than an independently collected original legal corpus.

Users of this dataset should cite both the original dataset and the associated paper when using the Bodo data in research.

Original Sources

• —Zenodo dataset: https://doi.org/10.5281/zenodo.7152317
• —ACL Anthology paper: https://doi.org/10.18653/v1/2022.aacl-main.77

Dataset Splits

The documents were split at the document level using a fixed random seed (42), so a judgment and its corresponding summary always remain in the same split.

Intended Uses

The dataset may be useful for:

• —Bodo legal text summarization
• —Abstractive summarization
• —Low-resource NLP research
• —Bodo language model fine-tuning
• —Legal NLP
• —Cross-lingual and multilingual summarization research

Data Format

The dataset is provided in JSON Lines (JSONL) format, with one judgment-summary pair per line.

Example:

{
  "id": "001",
  "language": "Bodo",
  "domain": "Legal",
  "judgment": "Full Bodo legal judgment...",
  "summary": "Corresponding Bodo summary..."
}

Loading the Dataset

from datasets import load_dataset

dataset = load_dataset("YOUR_USERNAME/bodo-legal-judgment-summarization")

print(dataset)
print(dataset["train"][0])

Citation

If you use this Bodo dataset, please cite the original work:

@inproceedings{shukla-etal-2022-legal,
    title = "Legal Case Document Summarization: Extractive and Abstractive Methods and their Evaluation",
    author = "Shukla, Abhay and
              Bhattacharya, Paheli and
              Poddar, Soham and
              Mukherjee, Rajdeep and
              Ghosh, Kripabandhu and
              Goyal, Pawan and
              Ghosh, Saptarshi",
    booktitle = "Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing",
    pages = "1048--1064",
    year = "2022",
    publisher = "Association for Computational Linguistics",
    doi = "10.18653/v1/2022.aacl-main.77"
}

Licensing and Redistribution

The appropriate license for this derived dataset should be selected only after confirming the redistribution rights and terms applicable to the original source material and the Bodo translations. Do not assume that translation alone creates unrestricted redistribution rights.

Users should consult the original Zenodo record and the applicable source terms before redistributing the underlying legal documents.

Acknowledgement

This resource was prepared to support research and development of Bodo-language NLP resources, particularly for legal text summarization. Credit is given to the original dataset creators and authors listed above, whose work provided the source material for this Bodo translation/adaptation.