View source for KREDIT: Knowledge infoRmEd NLU for Deception IdenTification


Accurate and consistent deception detection 669,837 human analysis and complete automation is likely impossible. We plan to help scale deception detection with assistive Artificial Intelligence (AI) technology. Our approach is to aid interrogators/interviewers to detect possible indicators of deception, and then prompt them to ask relevant follow-up questions to clarify or uncover the deception. Building upon an understanding of deception rooted in psychology, linguistics, and social sciences, we note that a binary classification problem based on natural language processing (NLP) is misguided and unhelpful. Instead, we propose to adopt methods termed knowledge-infused learning (KiL), which is a neurosymbolic AI approach that we have developed for natural language understanding (NLU). Rather than relying on training with large data corpora, KiL also uses a variety of knowledge sources (that humans use) to identify potential contributors of deception. Unlike the black-box nature of popular deep learning methods, we provide interpretability and explainability necessary to further tune AI algorithms for deceptive language. We envision three iterations of a basic deception detection pipeline, accruing increasing deception capabilities. First, we will develop proof-ofconcept deception detection algorithms that use syntactic, semantic, and pragmatic features to flag and explain suspicious text episodes to the interviewer. We identify the lack of a corpus as a major challenge and propose to develop a more useful and comprehensive corpus to evaluate the accuracy and effectiveness of our methods. Second, we refine this pipeline, tempering our techniques to take into account the base rate of deception in unbalanced corpora, which is essential to avoiding the false alarm problems that plague usability. Third, we expand the detection of deception heuristics that exploit information structure without being technically untruthful. Our team of accomplished researchers from the AI Institute (AIISC) and Institute of Mind and Brain (IMB) at the University of South Carolina (USC) and consultants from Wright State University (WSU), covers necessary expertise across psychology, linguistics, and computer science, and has a history of past collaborations resulting in published research, useful tools and commercialization.



== Funding ==


* ''' Amount ''' : $669,837

== Personnel ==
* ''' PI ''' : [http://amit.aiisc.ai/ Amit Sheth] (AIISC ; Computer Science & Engineering)
* ''' Co-PI ''' : [https://sc.edu/study/colleges_schools/artsandsciences/psychology/our_people/directory/almor_amit.php Amit Almor] (Department of Psychology; Institute of Mind and Brain), [https://people.wright.edu/valerie.shalin/ Valerie Shalin] (Department of Psychology, Wright State University)
* ''' Key Contributor ''' : [https://www.linkedin.com/in/spencer-m-seals-0b92761b9/ Spencer Seals] (Department of Psychology, Wright State University)