Solving dynamic-input interpretation problems using the hypothesize-test-revise paradigm

The authors introduce a dynamic-input problem solver and contrast it with the traditional static approach. A domain-independent hypothesis revision strategy for solving dynamic-input interpretation problems is presented, and its reasoning techniques are compared with R. Reiter´s (1980) default logic. A novel problem solver architecture based on a truth maintenance system is described. Advantages of this approach include a domain-independent strategy for generation and revising hypotheses in a dynamic environment, and a clear separation between the referential level and the inferencing level in a default reasoning system