A generative framsework for modeling human behavior with an application to analysis of deceptive behavior

Deception is a message knowingly and intentionally transmitted with the intent to foster false beliefs or conclusions. This definition encompasses strategies to mislead, such as equivocation, ambiguity, evasiveness and outright falsification. There are differences between deception and lying, but they have been ignored by the majority of research. However, wherever two people communicate, deception is a reality. It is present in our everyday social and professional lives and its detection can be beneficial, not only to us individually but to our society as a whole. For example, accurate deception detection can aid law enforcement officers in solving a crime. It can also help border control agents to detect potentially dangerous individuals during routine screening interviews. In this paper we propose a generative approach for learning the underlying dynamic structure of behavioral patterns and we apply it for recognizing high level tasks such as the difficulty of a question asked in a show and analysis of deceptive behavior.