Research

Engineering Sciences

Title :

Development of Multilingual and Multimodal Models for Online Deception Detection to Empower Law-enforcement Agencies

Area of research :

Engineering Sciences

Principal Investigator :

Dr. Md Shad Akhtar, Indraprastha Institute Of Information Technology, Delhi

Timeline Start Year :

2022

Timeline End Year :

2025

Contact info :

Details

Executive Summary :

Deception is the act of lying, making false statements, or hiding facts to achieve a goal. It can have a positive intention or negative intention, and can have severe consequences in investigations, particularly in the judiciary and law-enforcing agencies. Deception analysis is crucial for identifying deceptive statements and preventing potential threats to security, law and order, social disharmony, and injustice. Various interrogation techniques have been adopted in many countries, but textual statements are the most important evidence against deception. Deceivers often use different techniques, such as lack of self-reference, different tenses, and answering questions with questions. Non-verbal signals, such as lack of confidence, reluctant voice, and emotion shift, provide critical information. These clues are difficult to process in real environments due to human expertise, which is costly and time-consuming. A proposed intelligent multilingual and multimodal deception analysis system aims to assist investigators in interrogation, targeting both police and in-court judicial interrogation with multilingual support. The system will analyze the interrogatee during live proceedings, including uttered statements, visual expressions, and acoustics signals, and alert the interrogator for unusual or suspicious activities. Human observations and the system's fine-grained information can improve the probability of an effective interrogation by discriminating between guilty and innocent individuals. The system will also provide automatic translation support for two languages, English and Hindi. The project does not undermine the expertise of human investigators, but aims to assist them in the effective investigation.

Co-PI:

Dr. Tanmoy Chakraborty, Indraprastha Institute Of Information Technology, Delhi-110020

Total Budget (INR):

39,01,320

Organizations involved