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

arXiv: https://arxiv.org/abs/2507.14619 · HTML: https://mlsanity.com/paper/2507.14619/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 17 references that are in the map.
- First posted on arXiv 19 Jul 2025.

## Builds on (most cited first, 10 of 10)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Efficient Natural Language Response Suggestion for Smart Reply](https://mlsanity.com/paper/1705.00652/) (2017)
- [Multi-Stage Document Ranking with BERT](https://mlsanity.com/paper/1910.14424/) (2019)
- [Learning Dense Representations for Entity Retrieval](https://mlsanity.com/paper/1909.10506/) (2019)
- [In Defense of Cross-Encoders for Zero-Shot Retrieval](https://mlsanity.com/paper/2212.06121/) (2022)
- [Legal Document Retrieval using Document Vector Embeddings and Deep Learning](https://mlsanity.com/paper/1805.10685/) (2018)
- [Attentive Deep Neural Networks for Legal Document Retrieval](https://mlsanity.com/paper/2212.13899/) (2022)
- [ACORD: An Expert-Annotated Retrieval Dataset for Legal Contract Drafting](https://mlsanity.com/paper/2501.06582/) (2025)
- [Enhancing Legal Document Retrieval: A Multi-Phase Approach with Large Language Models](https://mlsanity.com/paper/2403.18093/) (2024)

## Abstract (arXiv)

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.