Abstract

We describe a general problem solving mechanism that is especially suited for performing a particular form of abductive inference, or best-explanation finding. A problem solver embodying this mechanism synthesizes composite hypotheses. It does so by by combining hypothesis parts as a means to the satisfaction of explanatory goals. In this way it is able to arrive at complex, integrated conclusions which are not pre-stored.The intent is to present a computationally-feasible, task-specific problem solver for a particular information processing task which is nevertheless of very great generality. The task is that of synthesizing coherent composite explanatory hypotheses based upon a prestored, and possibly vast collection of hypothesis-generating “concepts”. The authors’ claim is nothing less than to have shown, in a new sense, and surpassing all other work in this area, how it is computationally possible for an agent to come to “know”, based upon the evidence of the case.