Evaluating Human Eye Features for Objective Measure of Working Memory Capacity

With the advances in large language models such as GPT-3 and Megatron-Turin NLG, artificial intelligence systems are capable of synthetically generating misinformation. Wide use of these AI-ge- generated texts and their capability to mimic human-created news aid the spreading of online misinformation. Understanding the behavioral and physiological factors of humans when consuming fake information can be useful in detecting fake news. The advancement of eye-tracking technology allows us to use eye movement to assess human visual attention, including cognitive load measures.

In this talk, Yasasi Abeysinghe (Ph.D.) will discuss how we can gain insight into human behaviors in relation to the truthfulness of news and their believability using eye movements. She will present their analysis of advanced eye gaze measures to find relationships between visual search dynamics and cognitive load with respect to truthfulness and perceived believability of news content.

Yasasi Abeysinghe, PhD
B.Sc. Engineering (Hons) degree in Computer Science and Engineering at the University of Moratuwa, Senior Software Engineer at Sysco LABS, Ph.D. Candidate at Old Dominion University, Norfolk, Virginia, majoring in Computer Science with a focus on neuro information retrieval and data science.

The research group SPiRIT has been cooperating with the Computer Science department of Old Dominion University since 2012. Current projects are related to the @WebSciDL group of Professors Michael L. Nelson and Michele C. Weigle and the NIRDs group of Sampath Jayarathna.

This presentation was made possible by the projects h2d2 and ZAKKI at Magdeburg-Stendal University.