Reasoning in paraconsistent frame systems

Knowledge representation is of crucial importance in the held of Artificial Intelligence. Though there is no general consensus on what is knowledge representation, many schemes were proposed to represent and store knowledge. This paper focuses on frames. Frames are a knowledge representation scheme allowing the description of complex objects. However, a gap exists between the knowledge represented by frame systems and knowledge in the real world. Few frame systems are concerned with bridging this knowledge gap; they do not handle adequately issues such as exceptions and the inconsistency phenomenon. One of the shortcomings found in such systems is the lack of a formal semantics to treat exceptions and inconsistencies. In this paper a paraconsistent inheritance reasoner in Paralog_e is implemented, representing knowledge by means of paraconsistent frames and making inferences on tangled hierarchies, based on the degree of inconsistency/under-determinedness. Furthermore, its main characteristic is not eliminating contradictions ab initio. The inheritance reasoner implemented allows more adequate treatment of exceptions and inconsistent information in multiple inheritance frame systems.