Deception analysis: controlled dialogue summarization

Abstract

In the field of deception analysis, controlled dialogue summarization plays a critical role in dis tilling key information from witness testimonies. This project aims to develop a summarization framework for interrogations related to various crimes. The dataset comprises interrogations of witnesses who have been shown different videos related to crimes. The proposed system aims to produce comprehensive summaries, and extracts key information based on attributes such as Event Details, Factual information, Character Description ,etc. This work has significant implications for legal domains, facilitating better understanding and evaluation of testimonies in criminal investigations.