Deception Detection and Analysis in Spoken Dialogues based on FastText

Detecting deception is complicated for humans even though it often happens in human communications. In contrast, machines can capture small features to achieve accurate deception-detection, which is difficult for humans. Classifiers based on supervised learning make it possible to analyze effective features for deception-detection by giving positive and negative samples of deception to the classifier. FastText is one accurate classifier for a variety of classification problems, sentiment analysis, or the tagging of sentences, all of which use the distributed representation of features. We constructed a deception detector for dialogue utterances by giving labels of deception to FastText. We also combined acoustic features for deception-detection and analyzed the deception-detection results. The resultant detector achieved significantly higher accuracy than deception-detection by humans.

[1]  Andreas Stolcke,et al.  Distinguishing deceptive from non-deceptive speech , 2005, INTERSPEECH.

[2]  Tomas Mikolov,et al.  Bag of Tricks for Efficient Text Classification , 2016, EACL.

[3]  Robert S. Feldman,et al.  Looking Good and Lying to Do It: Deception as an Impression Management Strategy in Job Interviews , 2006 .

[4]  Sherali Zeadally,et al.  Multiple Account Identity Deception Detection in Social Media Using Nonverbal Behavior , 2014, IEEE Transactions on Information Forensics and Security.

[5]  B. Depaulo,et al.  Accuracy of Deception Judgments , 2006, Personality and social psychology review : an official journal of the Society for Personality and Social Psychology, Inc.

[6]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[7]  Elliot Moore,et al.  A study of Glottal waveform features for deceptive speech classification , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[8]  Eileen Fitzpatrick,et al.  Verification and Implementation of Language-Based Deception Indicators in Civil and Criminal Narratives , 2008, COLING.

[9]  Julia Hirschberg,et al.  Cross-Cultural Production and Detection of Deception from Speech , 2015, WMDD@ICMI.

[10]  M. A. Campion,et al.  Measuring faking in the employment interview: development and validation of an interview faking behavior scale. , 2007, The Journal of applied psychology.

[11]  Björn Schuller,et al.  Opensmile: the munich versatile and fast open-source audio feature extractor , 2010, ACM Multimedia.

[12]  Li Shang,et al.  Deception detecting from speech signal using relevance vector machine and non-linear dynamics features , 2015, Neurocomputing.

[13]  Erik Marchi,et al.  Is Deception Emotional? An Emotion-Driven Predictive Approach , 2016, INTERSPEECH.

[14]  W. P. Robinson,et al.  Truth, Equivocation Concealment, and Lies in Job Applications and Doctor-Patient Communication , 1998 .

[15]  Malcolm R. Parks,et al.  Deception Detection and Relationship Development: The Other Side of Trust , 1986 .

[16]  Timothy R. Levine,et al.  Deception Detection Accuracy is a Predictable Linear Function of Message Veracity Base-Rate: A Formal Test of Park and Levine's Probability Model , 2006 .

[17]  Jeffrey Pennington,et al.  GloVe: Global Vectors for Word Representation , 2014, EMNLP.

[18]  Julia Hirschberg,et al.  Hybrid Acoustic-Lexical Deep Learning Approach for Deception Detection , 2017, INTERSPEECH.

[19]  Mohamed Abouelenien,et al.  Verbal and Nonverbal Clues for Real-life Deception Detection , 2015, EMNLP.

[20]  Björn W. Schuller,et al.  The INTERSPEECH 2009 emotion challenge , 2009, INTERSPEECH.

[21]  F ROSENBLATT,et al.  The perceptron: a probabilistic model for information storage and organization in the brain. , 1958, Psychological review.