On Context and the Open World Assumption

Context-aware systems have historically employed a limited range of available data and while research across a diverse range of domains is addressing an increasingly broad range of data the level of intelligence generated in context-aware systems remains generally very limited. This is arguably due to the highly dynamic nature and inherent complexity of context. Individuals view the world though a unique perceptual filter developed over time, such a filter develops and changes based on an individuals experience of the environment in which they exist and contacts made in the course of social interactions. An individual's reaction to stimuli is driven by their view of the world and an important reactive component is emotion (more accurately emotional response). This paper considers context and the 'next-generation' intelligent context-aware systems incorporating decision-support with emotional response. Traditional approaches to knowledge have adopted the closed world assumption however we argue that adoption of an open world assumption offers exciting potential for improving the levels of intelligence in context-aware systems. The challenges in adopting the open world assumption are discussed with potential solutions and future directions for research. There are exciting opportunities for achieving improved levels of personalisation in context-aware systems if context processing predicated on the Open World Assumption can be realised.