a) On the one side, because it is the intelligence necessary for an entity that acts in an open world, interfering/interacting (either by competition or by cooperation) with other active entities (MAS; social ALife); or for systems able to really collaborate and to dialog with the user (new social paradigm in interface and in HCI); or for systems able to efficaciously support human interaction, cooperation, and virtual societies (CSCW, computer mediated interaction, etc.).
b) On the other side, "social" in the sense of "collective", i.e. that intelligence and problem-solving ability spontaneously emerging from the interaction of (more or less intelligent) entities which do not resolve a problem individually and thanks to their understanding. This "social" paradigm goes even beyond AI. It is some sort of more general view of computation, and of HCI.
Also, this new paradigm, that has been mainly developed in the DAI, MAS and Agents communities, lacks what we would call its "conceptual foundations". We mean that the foundational notions and perspectives of artificial social intelligence are vague, badly defined, contradictory, non systematic, etc. Or -even worst- people acritically import pieces of theories and notions from the social sciences, as mere instruments (for ex. game theory) without understanding the very serious theoretical problems and discussions which lie behind those concepts/theories. On the one hand, good formal instruments (logics, decision theory, etc.) cannot fill this gap. A good formalisation of a bad ontology cannot provide a good basis for a scientific enterprise or for practical systems. We claim that conceptual foundations are not less important for AI than formal foundations. On the other hand, this passive and importing attitude by AI bears troublesome implications, not only for its scientific status but also for its possible, necessary and revolutionary contribution to the social sciences. We claim that, as in the '60s and '70s the cultural influence of AI gave rise to Cognitive Science, now it will strongly impact on the social sciences.
This tutorial is aimed at giving a contribution to these processes by focusing of some of those foundational cognitive notions and their interesting consequences for Agents and MAS.
Aims
The Course is aimed at providing -on the basis of the work of the MAS and Agents communities and of the IP-CNR group- some of the most important conceptual foundations of Agents and their interaction and social intelligence (escluding the point in (b)). In particular, (most of) the following topics will be addressed (also depending on the interests of the audience and the discussion):
