AI enabled legal assistance system: A case study

Given a Case, finding the related prior cases and their judgements is a time consuming job of a Lawyer. The lawyer has to go through huge volumes of law books and prepare his case.  An automated tool that retrieves the relevant past cases and their judgments is a very useful application for lawyers.  It is a complex task especially in Indian context, the cases and their judgments are un-structured, and there is no standard format of case and judgement presentation.  It is understandable for a lawyer but, a most difficult task for a machine. The work here presents a case study to retrieve judgments given in the past for a given factual description. The dataset considered for this work is selected from FIRE-2019, AILA track.  The previously developed models showed best average precision of 0.149 using BM25, which itself demonstrates the challenging aspect of the given task.   In this work LDA, a probabilistic algorithm for Topic Modelling is explored and studied. The proposed method has shown improved precision