Abstract: This paper proposes social computation, i.e. large-scale man-machine collaboration mediated by digital interaction media, as a vision for future intelligent systems, and as a new challenge for multiagent systems research. We claim that the study of social computation suggests a re-interpretation of many traditional AI endeavours, has huge potential application benefits, and presents the field of multiagent systems with novel, exciting research questions. We introduce an abstract model of social computation that helps capture some of its core research problems more precisely. We explore the potential contribution of multiagent systems technologies to the solution of these problems by exposing the close relationship between social computation and existing methods in multiagent systems. We describe how these methods could be reused in this novel application context, what methodological implications this has, and argue that the resulting cross-fertilisation will be highly beneficial for both sides.
Keywords: social computation, human-based computation, crowdsourcing, collective intelligence.
Citation: M. Rovatsos. Multiagent Systems for Social Computation (challenge paper), Thirteenth International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2014), May 5-9, 2014.