Deep Personality Recognition For Deception Detection

Researchers in both psychology and computer science have suggested that modeling individual differences may improve the performance of automatic deception detection systems. In this study, we fuse a personality classification task with a deception classifier and evaluate various ways to combine the two tasks, either as a single network with shared layers, or by feeding personality labels into the deception classifier. We show that including personality recognition improves the performance of deception detection on the Columbia X-Cultural Deception (CXD) corpus by more than 6% relative, achieving new state-of-the-art results on classification of phrase-like units in this corpus.