[Next] [Previous] [Up] [Top] [Contents]
6. Bottom-up and Top-down Approaches to Modelling Context
6.1. Top-down
The approach to modelling context most frequently taken in AI, and perhaps epitomised by the approach of John McCarthy [12], which is to specify a general structure for representing statements concerning contextual reasoning and then to investigate some of the possible axiomatisations of logics that encapsulate principles that are thought desirable. Thus the general principles are formulated first and the properties emerging from these are investigated later. The initial standard for judging such constructs is the plausibility and generality of the abstract principles - thus like mathematics this is a foundationalist approach. Of course, the ultimate judgement comes from the usefulness of the approach in formalising or implementing actual systems. This approach is partly a result of a desire to elucidate generally applicable AI principles and partly a bias resulting from the selection of formal logics as a tool for modelling practical reasoning.
Of course, some work in AI takes a less general approach than this, especially where the work is focused towards a specific problem or problem domain.
The Pragmatic Roots of Context - Bruce Edmonds - 31 MAR 99
[Next] [Previous] [Up] [Top] [Contents]
Generated with CERN WebMaker