Information Filters and Their Implementation in the SYLLOG System

Abstract

This chapter presents a general framework for the reduction of the harmfulness of learned knowledge. The study of knowledge has always been one of the central issues of the AI research. The potential harmfulness of correct knowledge has become a prominent concern alongside the harm because of incorrect knowledge. Knowledge is harmful if the costs associated with retaining it are greater than its benefits. Irrelevant knowledge and redundant knowledge are two types of knowledge that are very often harmful. When the knowledge is acquired by a learning program, it is desirable that the harmfulness of the knowledge will be eliminated or at least reduced by the program itself. Information in a learning system flows from the experiences that the system is facing, through the acquisition procedure to the knowledge base, and thence to the problem solver. An information filter is any process that removes information at any stage of this flow. The filters that are inserted between the experience space and the acquisition procedure data filters, and the filters that are inserted between the acquisition procedure and the problem solver knowledge filters. Information filters essentially are functions that are inserted between the input to the learning system and the input to the problem solver. The role of these functions is to eliminate (or reduce) harmful knowledge.