Despite the challenges of social distancing in 2020, the Scottish Graduate School of Social Science successfully pivoted online to deliver a full programme of events, including a tutorial by SOCIAL AI CDT.
‘Social AI: Endowing Machines with Social Intelligence through Artificial Intelligence Technologies’ was presented in three parts by the directors, Professor Alessandro Vinciarelli, Dr Monika Harvey and Professor Stacy Marsella.
Professor Vinciarelli begins with an overview of the aims of SOCIAL AI CDT and explains how artificial intelligence technologies contribute to the CDT’s four major research areas into the interaction between people and artificial agents (robots, conversational interfaces, chatbots, virtual agents, etc.). The three themes of the tutorial are broadly ‘perception’, ‘cognition’ and ‘action’.
Professor Vinciarelli begins with the mapping of social perception of physical and inferential behavioural cues onto feature vectors for future learning by artificially intelligent agents.
Next, Dr Harvey provides an introduction to understanding the brain mechanisms underpinning perception and action, the general implications for perceiving and interacting with social agents, and ends with a focus on social agents interacting with special populations (e.g. stroke populations).
Finally, Professor Marsella outlines how human behaviour can be modelled computationally and then integrated into a virtual human for applications as diverse as entertainment and medical triage.