Two new publications on Social Situation Awareness of Personal Assistant Agents have been accepted in the past months (Ilir Kola as first author)!
One is an exploratory paper presented at the workshop on Modelling and Reasoning in Context. It investigates if (social) situations can be clustered based on so-called psychological characteristics of situations, and if these clusters can be linked to values that are promoted or demoted in these situations. The idea is that if a personal assistant agent could determine which values are affected by certain situations, it could provide support to its user that is more in alignment with their values.
The second is a follow-up and consolidation presented at the PRIMA’20 conference of an earlier feasibility study on the prediction of priorities of social situations based on their social features. This is relevant, for example for personal assistant agents that provide support in managing a user’s calendar: if it knows which social setting such as a meeting is more important for the user, they can for example prioritise meetings accordingly. The more general underlying question of the work concerns whether we could predict what a social situation means to the user on the basis of information about that situation, in particular how the user relates to people in that situation.
Check these papers out or contact us if you want to know more!