SoRAIM Winter School — Understanding Human Behavior for Mental Wellbeing Robotic Coaches
The last plenary session opened the discussion on the last day of SoRAIM. The topic of the session was “Understanding Human Behavior for Mental Wellbeing Robotic Coaches” and consisted of an Invited talk from Prof. Hatice Gunes (https://www.cl.cam.ac.uk/~hg410/, University of Cambridge) on “Robotic Coaches for Mental Wellbeing: From the Lab to the Real World”, followed up by a presentation of the achievements of the SPRING project on “Multi-Modal Human Behaviour Understanding” by Dr. Lorenzo Vaquero Otal (https://citius.gal/team/lorenzo-vaquero-otal/, University of Trento).
Prof. Hatice Gunes motivated the need to take care of a worldwide health problem, which is mental health. In particular, on promoting practices for increasing mental well-being. Prof. Gunes described her lab’s approach to a mental wellbeing robotic coach starting from results on teleoperated robots. She also described the so-called novelty effect, which happens not only for robotic coaches but also for human ones, perhaps with different temporal spans. Prof. Gunes also presented results suggesting that the impact of the robotic wellbeing coach depends on the user’s personality, and the on-fits-all solution does not seem to be the most appropriate. In this regard, one possibility is to turn to the continual learning paradigm, and Prof. Gunes explained their work on personalised multimodal affect recognition via continual learning.
Following up, Dr. Lorenzo Vaquero Otal provided an overview of SPRING’s WP4 about understanding human behavior with multimodal data. After briefly discussing the prominent contributions to individual and group behavior recognition as well as affect and robot acceptance recognition, Dr. Vaquero Otal delved into details on their work on targetted gaze recognition, combining depth, gaze estimation, and object detection in a holistic framework.