RI: Nonpropositional Nonmonotonic Languages for Knowledge Representation

点击翻译按钮获取中文摘要

ABSTRACTKnowledge representation formalisms are precisely defined languages designed for expressing declarative knowledge (assertions), just as typical programming languages are used to express procedural knowledge (algorithms). Because of the important role of declarative knowledge in intelligent behavior, the theory of knowledge representation is a key part of artificial intelligence. Nonmonotonic knowledge representation languages are particularly valuable in view of the fact that they allow reasoning about defaults and exceptions. Research on nonmonotonic knowledge representation is becoming increasingly experimental and applied, in connection with the emergence of efficient answer set solvers--software systems for computing stable models.The goal of this project is to reformulate the semantics of variables in nonmonotonic logic in a way that will bring this theory closer to the reality of state-of-the-art implementations. This will be achieved using a new definition of a stable model that is based on translating logic programs with variables into classical logic. This work seeks to clarify and simplify the semantics of several knowledge representation languages that are used in applications of artificial intelligence to many areas of science and technology.