Generative Chinese Statute Retrieval

2026-07-13

AI Summary

This study addresses the semantic gap between everyday legal queries and formal legal provisions by introducing generative retrieval to Chinese legal article retrieval for the first time. The authors propose a sequence-generation-based retrieval framework that internalizes legal knowledge through multi-task training and designs multi-granularity structured document identifiers (docids) incorporating both the hierarchical structure of legal codes and semantic information. Experimental results demonstrate that the proposed model significantly outperforms sparse, dense, and legal-domain-specific baselines across multiple evaluation metrics, confirming the effectiveness and potential of generative approaches for legal information retrieval and downstream reasoning tasks.

Abstract

Statute retrieval is a fundamental task in legal information retrieval, yet existing approaches struggle to bridge the gap between colloquial legal queries and formal statutory language. In this paper, we propose GCSR, a generative statute retrieval framework that reformulates statute retrieval as a sequence generation problem and internalizes statutory knowledge into a generative model. Specifically, we propose a multi-granularity structured docid that encodes legal hierarchy and semantic information, together with a multi-task training strategy. Experiments show that GCSR consistently outperforms strong sparse, dense, and legal-domain baselines. Our results demonstrate the effectiveness of generative retrieval for statute retrieval and highlight its potential for broader legal information access and downstream legal reasoning tasks.

Problem

• statute retrieval
• legal information retrieval
• colloquial legal queries
• formal statutory language
• generative retrieval

Innovation

• generative retrieval
• statute retrieval
• structured docid
• multi-task training
• legal information retrieval