Invited Speaker: Julia Hirschberg, Columbia University Truth or Lie? Spoken Indicators of Deception in Speech

摘要

Detecting deception from various forms of human behavior is a longstanding research goal which is of considerable interest to the military, law enforcement, corporate security, social services and mental health workers. However, both humans and polygraphs are very poor at this task. We describe more accurate methods we have developed to detect deception automatically from spoken language. Our classifiers are trained on the largest cleanly recorded corpus of within-subject deceptive and non-deceptive speech that has been collected. To distinguish truth from lie we make use of acoustic-prosodic, lexical, demographic, and personality features. We further examine differences in deceptive behavior based upon gender, personality, and native language (Mandarin Chinese vs. English), comparing our systems to human performance. We extend our studies to identify cues in trusted speech vs. mistrusted speech and how these features differ by speaker and by listener. Why does a listener believe a lie?