Social machines for education driven by feedback agents

The aim of this paper is to explain some of the ways in which multi agent system (MAS) theory can be used to describe, design and enhance social machines (also referred to as Socio-Cognitive Systems). We believe there is a really opportunity for the MAS community to engage with emerging theory and practice of designing such systems. Social machines also referred to as Socio-Cognitive Systems from the MAS community are terms used to refer to the recent breed of technological systems which allow human and computational agents to socially interact, typically on a large scale and sometimes towards achieving shared goals. Examples include social networking platforms and crowd sourced encyclopaedias. The discussion of social machines and MAS is taken from three perspectives. Firstly, the theoretical notion of an abstract social machine as a socio-cognitive system containing humans and agents is introduced. Secondly, a specific instance of a social machine which has been designed to enable social music learning supported by agents is described. Thirdly, an agent architecture which is designed for operation within educational social machines is discussed, with particular focus on what we believe is the core currency of these machines: feedback.