Optimizing Legal Document Retrieval in Vietnamese with Semi-Hard Negative Mining

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Le, Van-Hoang, Nguyen, Duc-Vu, Van Nguyen, Kiet, Nguyen, Ngan Luu-Thuy
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916852070875136
author Le, Van-Hoang
Nguyen, Duc-Vu
Van Nguyen, Kiet
Nguyen, Ngan Luu-Thuy
author_facet Le, Van-Hoang
Nguyen, Duc-Vu
Van Nguyen, Kiet
Nguyen, Ngan Luu-Thuy
contents Large Language Models (LLMs) face significant challenges in specialized domains like law, where precision and domain-specific knowledge are critical. This paper presents a streamlined two-stage framework consisting of Retrieval and Re-ranking to enhance legal document retrieval efficiency and accuracy. Our approach employs a fine-tuned Bi-Encoder for rapid candidate retrieval, followed by a Cross-Encoder for precise re-ranking, both optimized through strategic negative example mining. Key innovations include the introduction of the Exist@m metric to evaluate retrieval effectiveness and the use of semi-hard negatives to mitigate training bias, which significantly improved re-ranking performance. Evaluated on the SoICT Hackathon 2024 for Legal Document Retrieval, our team, 4Huiter, achieved a top-three position. While top-performing teams employed ensemble models and iterative self-training on large bge-m3 architectures, our lightweight, single-pass approach offered a competitive alternative with far fewer parameters. The framework demonstrates that optimized data processing, tailored loss functions, and balanced negative sampling are pivotal for building robust retrieval-augmented systems in legal contexts.
format Preprint
id arxiv_https___arxiv_org_abs_2507_14619
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Optimizing Legal Document Retrieval in Vietnamese with Semi-Hard Negative Mining
Le, Van-Hoang
Nguyen, Duc-Vu
Van Nguyen, Kiet
Nguyen, Ngan Luu-Thuy
Information Retrieval
Computation and Language
Large Language Models (LLMs) face significant challenges in specialized domains like law, where precision and domain-specific knowledge are critical. This paper presents a streamlined two-stage framework consisting of Retrieval and Re-ranking to enhance legal document retrieval efficiency and accuracy. Our approach employs a fine-tuned Bi-Encoder for rapid candidate retrieval, followed by a Cross-Encoder for precise re-ranking, both optimized through strategic negative example mining. Key innovations include the introduction of the Exist@m metric to evaluate retrieval effectiveness and the use of semi-hard negatives to mitigate training bias, which significantly improved re-ranking performance. Evaluated on the SoICT Hackathon 2024 for Legal Document Retrieval, our team, 4Huiter, achieved a top-three position. While top-performing teams employed ensemble models and iterative self-training on large bge-m3 architectures, our lightweight, single-pass approach offered a competitive alternative with far fewer parameters. The framework demonstrates that optimized data processing, tailored loss functions, and balanced negative sampling are pivotal for building robust retrieval-augmented systems in legal contexts.
title Optimizing Legal Document Retrieval in Vietnamese with Semi-Hard Negative Mining
topic Information Retrieval
Computation and Language
url https://arxiv.org/abs/2507.14619