SOME REPRESENTATIONAL ISSUES IN DEFAULT REASONING

Network representations fail to reason correctly with defaults, but when viewed as indexing schemes on logical formulae, networks can be seen to provide computationally feasible heuristics for the consistency checks required by default reasoning.

Abstract

Although most commonly occurring default rules are normal when viewed in isolation, they can interact with each other in ways that lead to the derivation of anomalous default assumptions. In order to deal with such anomalies it is necessary to re-represent these rules, in some cases by introducing non-normal defaults. The need to consider such potential interactions leads to a new concept of integrity, distinct from the conventional integrity issues of first order data bases. The non-normal default rules required to deal with default interactions all have a common pattern. Default theories conforming to this pattern are considerably more complex than normal default theories. For example, they need not have extensions, and they lack the property of semi-monotonicity. Current semantic network representations fail to reason correctly with defaults. However, when viewed as indexing schemes on logical formulae, networks can be seen to provide computationally feasible heuristics for the consistency checks required by default reasoning.