Speedup of Hypothetical Reasoning Based on Run-Time Knowledge Reformation

Abstract

A logic-based hypothetical reasoning system, which employs the mechanism of abduction, is a useful knowledge processing framework applicable to many practical problems including diagnosis, design, etc. However, it has a crucial problem of slow inference-speed due to its non-monotonic inference nature. Knowledge-base reformation, which has been executed so far before the real usage of the knowledge-base, is one way to overcome this problem. This paper presents a run-time knowledge reformation method, which works as a prepossessing connected to networked methods of hypothetical reasoning. This reformation method consists of three phases. The first phase is the partial compilation of extracted knowledge as relevant one to the proof of a given goal. The second phase is to reform the knowledge structure for removing its redundancy. The third phase is the transformation of the compiled knowledge into multi-level logical knowledge so as to remove further redundancy of the knowledge structure. The speed of two hypothetical reasoning methods has been improved, i. e., the total cpu-time including both the reforming process and their own reasoning computation has been remarkably reduced to a large extent particularly when original knowledge structure is complex for hypothetical reasoning task.