Representation and inference with consistent temporal propositions

Abstract

Reasoning with temporal information is essential in Artificial Intelligence. We consider a knowledge base where the internal representation language deals with temporally qualified propositions and constraints on the ordering of time points. As temporal information is typically partial, a representation including constraints on the order of temporal objects is particularly suited. Temporal statements associate maximal intervals to basic atemporal propositions. Queries posed to a knowledge base which includes facts and rules use deduction to explore its consequences and abduction to generate consistent hypotheses. The inference system relies on a set of constraint primitives providing temporal consistency both for points and for intervals. The relevant aspects of the temporal framework are the underlying propositional language, the abductive derivation procedure and the facility of built-in constraint handling. The abductive procedure in the inference system provides the strategy for completing partial information in order to produce informative answers. The integration of constraint solving with abduction allows a query oriented generation of answers where the enforcement of constraints is efficiently performed.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.