Mwnthai/bodo-legal-question-answering-iiith

Bodo Legal Question Answering Dataset — IIITH Translation

Overview

A Bodo-language legal Question Answering (QA) resource derived from English legal judgments. Each example contains a judgment context, a question, and its corresponding answer.

Data Provenance

Original Legal Source

The underlying English legal judgments were extracted from the publicly accessible Gauhati High Court judgment repository:

https://ghconline.gov.in/index.php/judgments-up-to-2017/

Bodo Translation

The English legal material was translated into Bodo using the IIITH translation model/service associated with the IIIT-H Language Technologies Research Centre's SSMT platform:

https://ssmt.iiit.ac.in/

The translation methodology/model is described in:

IIITH translation model paper: https://arxiv.org/pdf/2412.04351

This is therefore a derived, machine-translated Bodo resource based on publicly accessible Gauhati High Court judgment material.

Dataset Construction

1. 1.English legal judgments were extracted from the Gauhati High Court public judgment resource.
2. 2.The English text was translated into Bodo using the IIITH translation model/service.
3. 3.Question-answer data was constructed from the translated legal contexts.
4. 4.Each question-answer pair is represented as a separate example while retaining its source judgment context.

Dataset Statistics

• —Training source documents: 3,982
• —Test source documents: 820
• —Training QA pairs: 19,550
• —Test QA pairs: 4,000
• —Total QA pairs: 23,550
• —Source records without QA pairs: 92

The supplied train/test separation has been retained.

Dataset Fields

• —id: Unique example identifier.
• —title: Source judgment filename.
• —language: Bodo.
• —domain: Legal.
• —context: Bodo legal judgment context.
• —question: Bodo question.
• —answers.text: Answer text.
• —answers.answer_start: Character position of the answer in the context.

Loading

from datasets import load_dataset

dataset = load_dataset("YOUR_USERNAME/bodo-legal-question-answering-iiith")
print(dataset)
print(dataset["train"][0])

Intended Uses

• —Bodo Question Answering
• —Legal Question Answering
• —Extractive Question Answering
• —Machine Reading Comprehension
• —Low-resource NLP
• —Bodo language processing
• —Legal NLP
• —Cross-lingual and multilingual QA

Translation Quality

The Bodo text was generated through machine translation using the IIITH translation model/service. Machine translation may introduce linguistic, terminological, or contextual errors, especially in specialized legal content. This dataset should not be treated as an authoritative legal translation.

Attribution

Please acknowledge:

1. 1.Gauhati High Court for the underlying public judgments.
2. 2.IIITH / IIIT-H Language Technologies Research Centre for the translation model/service.
3. 3.The authors of the IIITH translation model paper.
4. 4.The creator(s) of this derived Bodo QA dataset.

Original source: https://ghconline.gov.in/index.php/judgments-up-to-2017/

IIITH translation platform: https://ssmt.iiit.ac.in/

IIITH model paper: https://arxiv.org/pdf/2412.04351

Disclaimer

This dataset is intended for research and educational purposes. It does not constitute legal advice or an authoritative translation or interpretation of any court judgment.

Licensing and Redistribution

Public accessibility of the judgments does not by itself establish unrestricted redistribution rights. Before making this dataset public, verify the applicable terms of use, copyright, and redistribution conditions of the original Gauhati High Court source.

Select a Hugging Face repository license only after confirming the rights applicable to the underlying judgments, Bodo translations, and QA annotations.

Citation

When publishing research using this resource, acknowledge the original Gauhati High Court source, the IIITH translation model paper and platform, and this dataset.