印度法律先例检索：词汇、语义与修辞角色信息研究
原标题：A Study of Prior Case Retrieval Using Lexical, Semantic, and Rhetorical Role Information in Indian Legal Documents
AI 摘要
该论文针对印度法律判决中的先例检索（IL-PCR）任务，实证评估了三种连续设计的检索系统。研究发现，结合修辞角色（事实、判决理由、先例、论点、法规）优于单独使用角色或全文检索，法规相似度不具区分性，而由LLM生成训练数据的法律蕴含重排序器在官方评估中远差于内部验证分数。作者提出采用查询不相交切分和冻结验证融合的新设计，最终四阶段流水线取得Micro F1 0.2549、MRR 0.6081、nDCG@10 0.4187。
正文节选
A Study of Prior Case Retrieval Using Lexical, Semantic, and Rhetorical Role Information in Indian Legal Documents Abstract For retrieving prior cases in Indian legal judgments, the problem involves distinguishing relevant legal facts from mere lexical similarities because a prior case that shares a statute with the query judgment is not necessarily relevant. In this paper, we describe an empirical evaluation of three consecutive designs of retrieval systems for the IL-PCR(Indian Legal Prior Cas