Evolving Learning in the Stuff Swamp
Jon Dron (University of Brighton, UK), Chris Boyne (University of Brighton, UK) and Richard Mitchell (University of Brighton, UK)
Copyright: © 2003
This chapter describes the theory, background and some uses of CoFIND (Collaborative Filter in N Dimensions), a Web-based database of learning resources which is created by and for learners. CoFIND is designed to exploit principles of evolution and self-organisation to create an emergent structure to learning resources. Through the manipulation of learner-supplied metadata such as classifications and ratings, this structure shapes itself to the needs of the learners who create it, providing something akin to guidance traditionally supplied by a teacher. The chapter starts with a discussion of the weaknesses of existing means for groups of learners to discover learning resources including search engines, directories, seals of approval, and collaborative filters. It considers a range of methods by which self-organisation is achieved in natural systems (notably evolution and stigmergy) and which underpin the CoFIND system. CoFIND is described and examples are given of some of its uses. The authors discuss some issues which arise, especially its cold-start problem, influences of surrounding systems and the role of motivation. The chapter concludes with a discussion of potential future directions for CoFIND and identifies some other aspects of learning environments which may benefit from such a self-organising system.