SBIR-STTR Award

Significant challenges exist related to smartly interviewing multiple people to assess eyewitness accounts and judge how truthful a source is. Aetion proposes applying a powerful form of inference called abduction to help intelligently interview multiple people and compare their corresponding answers to judge the reliability of a source. Abductive inference, or "inference to best explanation," has been intensely researched over the past twenty years. Aetion has successfully applied this approach to complex, real-world problems. Because abduction allows a high degree of abstraction, incongruous data is easily fused and intelligently analyzed. An abduction machine (software implementing an abductive inference approach) has the ability to assess the the truthfulness of a source, find potential relationships among sources, determine what statements need to be verified, and provide display support for the user.

Keywords:
Abductive Inference, Interview Tool, Hypothesis, Truth, Automated Interviewer