Communications of the ACM
Software That Listens for Lies
Dan Jurafsky of Stanford is among those who are teaching computers how to spot the patterns of emotional speechthe kind that reflect deception, friendliness and, yes, even flirtation.
Credit: Linda A. Cicero/Stanford News Service
Several linguists, engineers, and computer scientists are developing computer systems that can recognize signs of emotional speech, such as deception, anger, friendliness, and flirtation.
The technology is advancing quickly as labs share research, says Stanford University's Dan Jurafsky, who has been studying the language that people use in four-minute speed-dating sessions. "The scientific goal is to understand how our emotions are reflected in our speech, [and] the engineering goal is to build better systems that understand these emotions," Jurafsky says.
Columbia University professor Julia Hirschberg has developed a system that can identify 70 percent of lies. The algorithms are based on an analysis of the different ways people spoke in a research project when they lied or told the truth.
Meanwhile, University of Southern California professor Shrikanth Narayanan is using computers to analyze emotional speech by focusing on hundreds of cues such as pitch, timing, and intensity to distinguish between patterns of angry and non-angry speech.
Montclair State University professor Eileen Fitzpatrick is using computers to identify groups of words that could signal deception. Stanford University professor David F. Larcker recently used Jurafsky's research to analyze the words of financial executives who made statements that were later disproved.
From New York Times
View Full Article
Abstracts Copyright © 2011 Information Inc. 
, Bethesda, Maryland, USA
No entries found