In the context of social psychology, deception is the act of lying while communicating with other people. Identifying this act can be useful in a variety of contexts, from airport security to police interrogations and trial courts. Many studies in the last decades have shown that liars “leak” behavioral cues, both verbal and non-verbal, that are unconscious and hard to control, and that these cues can be spotted and identified. This activity has always been performed by humans, sometimes with the aid of auxiliary tools, like the polygraph. But, recently, automated deception detection has gained momentum thanks to the advances in the computer vision and machine learning research fields, which aid in the analysis and discovery of behavioral patterns that a human would hardly discover.
We are currently interested in the analysis of deception in real-life contexts, in both verbal and written form, through the use of multimodal data (such as audio and video recordings and text transcriptions), verbal and non verbal features, and the application of machine learning techniques that can leverage and combine the different modalities.