Knowledge Table : An Approach to Speeding up the Search for Relational Information in Knowledge Base

概要

論文の詳細を見る
For efficient knowledge utilization, it is necessary to keep relationships between knowledge such as isa (or class-inclusion) relationships, hasa (or part-whole) relationships, and data dependencies. These relationships are useful for managing and manipulating knowledge to solve complex problems in connection with belief revision, default reasoning, multiple inheritance within networks, and so on. Unfortunately, keeping and using them can be ineffective and expensive when large numbers of relationships are involved. In this paper, a knowledge table is introduced as a structure that allows effective searching for relationships in knowledge bases. The knowledge table is a table-like representation for keeping relationships that precludes the need for depth-first searching. To create the table, we use logical bitwise computations for searching and for representing relationships. This greatly speeds up the search for the relationships and makes it easy to find loop structures among the relationships.
一般社団法人情報処理学会の論文
1991-02-10