Bilkent University
Department of Computer Engineering

---

CS 590/690 SEMINAR

 

Multimodal Fusion Approaches for Deceit Detection

 

Ali Azak
Master Student
(Supervisor: Assoc.Prof.Hamdi Dibeklioğlu)

Computer Engineering Department
Bilkent University

Abstract: In this talk, various multimodal fusion approaches for deceit detection have been explored, and our results have been presented. Fusion approaches in deceit detection task remains an open challenge due to the complexity of integrating heterogeneous modalities, the variability of deceptive cues across different contexts, and the lack of large, diverse datasets that capture real-world deception scenarios. To solve these issues a summarization idea introduced in fusion process to decrease the complexity and achieve better generalization. Different fusion techniques have been evaluated on both low stake and high stake scenarios. Extensive experiments with various backbone architectures and fusion approaches are presented.

 

DATE: March 17, Monday @ 13:40 Place: EA 502