Recognizing unnecessary inference
Dan Benanav · International Joint Conference on Artificial Intelligence · 1989
Intelligent reasoners sometimes draw conclusions that lack new or relevant information. Similarly, automated reasoning systems can produce formulas that are not necessary for the problem at hand. We concentrate on the problem of unnecessary inference in the context of resolution based systems. In such systems several strategies have been developed that allow for the deletion of clauses without sacrificing completeness. Unfortunately these strategies fail to recognize other frequently generated unnecessary formulas. We will present a generalized subsumption theorem that can be used to recognize such formulas and to develop new deletion methods which retain completeness.