Scholarlink

<p>The legal domain presents unique challenges for information extraction and reasoning due to the complex structure and specialized language of legal texts. To tackle these challenges, our team, CAPTAIN, utilizes advanced Large Language Models (LLMs) to enhance legal information processing for the COLIEE 2025 competition. We participate in four tasks: Legal Case Entailment (Task 2), Statute Law Retrieval (Task 3), Legal Textual Entailment (Task 4), and Legal Judgment Prediction for Japanese Tort Law (Pilot Task). Our approach leverages the interpretive capabilities of LLMs to analyze and summarize intricate legal documents, identify semantic connections between cases and relevant statutes, and perform contextual reasoning. By employing diverse prompting techniques, we effectively uncover implicit relationships between legal cases and corresponding statutes, improving both interpretability and accuracy. Experimental results highlight the effectiveness of our method, with our team securing first place in the Tort Prediction sub-task of the Pilot Task and second place in both the Legal Statute Law Retrieval and Rationale Extraction sub-tasks, demonstrating the strong potential of LLM-based approaches in legal AI.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Legal Text Processing and Structural Analysis with Large Language Models at COLIEE 2025

  • 期刊论文

  • 发表日期：

    2026年3月25日

查看全文
  • Dat Nguyen,
  • Minh-Phuong Nguyen,
  • Quang-Huy Chu,
  • Son T. Luu,
  • Nguyen-Hoang Chu,
  • Trung Vo,
  • Le-Minh Nguyen

摘要

The legal domain presents unique challenges for information extraction and reasoning due to the complex structure and specialized language of legal texts. To tackle these challenges, our team, CAPTAIN, utilizes advanced Large Language Models (LLMs) to enhance legal information processing for the COLIEE 2025 competition. We participate in four tasks: Legal Case Entailment (Task 2), Statute Law Retrieval (Task 3), Legal Textual Entailment (Task 4), and Legal Judgment Prediction for Japanese Tort Law (Pilot Task). Our approach leverages the interpretive capabilities of LLMs to analyze and summarize intricate legal documents, identify semantic connections between cases and relevant statutes, and perform contextual reasoning. By employing diverse prompting techniques, we effectively uncover implicit relationships between legal cases and corresponding statutes, improving both interpretability and accuracy. Experimental results highlight the effectiveness of our method, with our team securing first place in the Tort Prediction sub-task of the Pilot Task and second place in both the Legal Statute Law Retrieval and Rationale Extraction sub-tasks, demonstrating the strong potential of LLM-based approaches in legal AI.