WISSLR 2016: Western Interdisciplinary Student Symposium on Language Research
Friday March 4, 2016 @1:20-2:20 p.m.
Western University, Arts and Humanities Building, 2R23

Keynote Speaker – Dr. Victoria Rubin “DECEPTION and Lie Spotting with Text Analytics”

ABSTRACT: Dr. Victoria Rubin will explore and explain the many types of deception, how deception is studied in various disciplines, and which linguistic patterns help predict the presence of deception in written texts. Recently, the media has shown a strong interest in the idea of computers discerning truth from deception. This interest is possible because of recent advances in the research and development of automated deception detection.

Rubin places the study of deception at the intersection of several disciplines – interpersonal psychology, information, media and communication studies, law enforcement and linguistically-aware text analytics. Deception is defined in computer-mediated communication as a message knowingly and intentionally transmitted by a sender to foster a false belief or conclusion by the perceiver. With the ever-increasing use of computer-mediated communication in all aspects of modern life, deception can be potentially disruptive in everyday communication, information seeking, and the retrieval and use of information for decision making.

Rubin argues that although largely condemned by society, deception is widespread and often undetected, especially in electronic environments where credibility is difficult to ascertain due to the absence of many traditional cues, such as verifiable credentials or face-to-face contact. The need arises for text analysis tools capable of alerting users to potentially deceptive content. Such tools can also be seen as a kind of decision-support software that suggests when any form of digital messages may be straying from the truth.

Rubin affirms that automated deception detection has recently become an attainable and practical contribution in human language technologies. The core of these capabilities is based on linguistic observations, and our understanding of the linguistic markedness in the language used by deceivers, by contrast with the linguistic characteristic of the language of truth-tellers. Deceptive cues are numerous, often unreliable, and not universally agreed upon in the detection research community. However, certain broad generalization about such cues as linguistic predictors of deception apply across varieties of deception types and circumstances. The most consistent linguistic characteristics associated with deception include increased word count, use of third-person pronouns, and markers of reduced cognitive complexity. Such patterns are usually invisible and unmeasurable to people without specialized tools.

With great optimism for the improved accuracy and effectiveness of the methods in deception detection research, Rubin discusses this relatively new area in applied computational linguistics. The envisioned potential outcomes of these cutting-edge computational abilities suggest that programs that succeed at spotting lies may someday have many practical uses in detecting business fraud and online misrepresentation, as well as in police and security work.