Abstract

Collective intelligence (CI), that is the combining of behaviour, preferences, or ideas of a group of people to create novel insights, is a core ingredient of today’s internet. In many cases the application of CI techniques is the only way to locate the one item of useful information in a deluge of networked data. Moreover, the accelerating adoption of pervasive computing in its many different flavours, leads to massive increases in data production due to the automation of the authoring process. As a consequence, the role of CI techniques is becoming even more marked. It is also becoming possible to go beyond analyses of online behaviours and draw information from the real world, which is now under constant observation through pervasive computing systems. In this talk I will introduce our recent work on CI techniques designed specifically for pervasive computing. In particular, I will discuss a model that comprehensively captures the features of CI in a manner suitable for a wider range of pervasive computing systems, efficient data structures for representing this model, and a collection of algorithms for analysis, prediction and classification. I will conclude by reporting on results obtained through the application of these techniques to a variety of data sets.