Semantic Search and Summarization of Judgments Using Topic Modeling

Online legal document libraries, such as WorldLII, are indispensable tools for professionals to conduct research. We study how topic modeling techniques can be applied platforms facilitate searching of court judgments. Specifically, we improve search effectiveness by matching judgments queries at semantics level rather than keyword level. Also, design a system that summarizes retrieved judgment highlighting small number paragraphs semantically most relevant the user query. This summary serves two purposes: (1) It explains why machine finds user’s query, and (2) it helps quickly grasp salient points judgment, which significantly reduces amount time needed go through returned results. further enhance our integrating domain knowledge provided experts. The includes features aspects important given category Users then view judgement’s focusing on particular only. illustrate with evaluation experiment HKLII platform. results show methods highly effective.