Mark Levene, Southampton
How Machines Can Learn to Adapt Their Behaviour to Suit Individual Learners
Kevin Keenoy, Mark Levene and Sara de Freitas
This chapter discusses how the trails concept can be applied in providing personalised learning, with a focus on the use of trails in adaptive e-leaming systems which adjust their behaviour in order to better cater for individual learners' needs.
The first half of the chapter looks at some of the technological and theoretical issues posed by personalised trails: We present a generic architecture for personalisation which can potentially be deployed on a variety of hardware configurations, and then look at learner profiling, a key aspect of personalisation. Leamer profiles store the information about learners that is used to provide personalisation, so the structure and content of the learner profile used by a system has implications for the kinds of personalisation and support for trails that can be provided. In this context we see how Semantic Web technologies can potentially be used to integrate disparate user profile specifications.
In the second half of the chapter we look at the practicalities and practice of personalising trails in the context of e-leaming. After a brief general discussion of some techniques that can be used to provide personalisation we use examples of real systems to demonstrate the potential for providing personalised trail support that currently exists. As well as systems that provide trails for learners our examples include adaptive hypertext techniques which provide a kind of trail support through adaptive presentation; systems that provide adaptive feedback to learners and which put learners in direct control of the adaptivity; social navigation systems and mobile computing systems. We conclude with some thoughts about the future of this burgeoning research area.