Detecting Deception in Multiscale Activity Graphs.

摘要

The problem studied in this research is that of detecting deception by participants in large-scale activities. The primary means pursued was that of obtaining a better understanding of how to describe and predict the flow of information through large populations of disparate types of participants. Such an understanding can be used to see if the observed results of deceptive information match what would be expected were the deception to originate from suspected sources. It can also be used to estimate susceptibility of participants or populations to deceptive information originating from different sources.