Multilingual Deception Detection by Autonomous Agents

In this work we present the development of a multilingual deception detection model based on speech. In addition, we also develop a model that detects whether a statement will be perceived as a lie or not by human subjects. To this end, we developed a game for collecting a large scale and high quality labeled data-set in a controlled environments in English and Hebrew. We developed a model that can detect deception based only on a vocal statement from the participants of the experiment. The data-set will be released to the community.