M.S. Candidate: Elif Esmer
Program: Cognitive Science
Date: 29.11.2022 / 16:00
Place: B116

Abstract: The purpose of this study is to investigate if gaze behavior modeling on a robot avatar makes any difference in the engagement of human interlocutors. To that end, visual interaction patterns and impressions of participants during a one-on-one active conversation setting in the form of a mock-up job interview with an avatar agent have been analyzed. Experiments were conducted through eye-tracking methodology, questionnaires, and open-ended post-experimental evaluations. A robot avatar as an artificial agent programmed with pre-recorded speech and pre-planned gaze behavior, based on human gaze patterns, took the role of the human resource manager, in other words the interviewer. The eye-tracking data of the participants, namely interviewees, were collected and analyzed to investigate whether the human-like gaze behavior of the artificial agent had any effect on the gaze allocation of the human interlocutors. Results of the questionnaires and the TF-IDF analysis of the post-experimental evaluations were inspected to detect reflections of participants’ gaze behavior and further patterns for the perception of the avatar.