Conceptual Structure: Theory And Applications

For more than a half century, artificial intelligence (AI) has got gradually much attention to many subjects, and become an inter-discipline and frontal science subject increasingly. In last 10 years modern computer already can store plenty of information and carry out fast information processing. Moreover, software function and hardware implement has made great progress. This makes AI get further applications.Knowledge representation is the central topic in AI field. Whether problem solving, or task describing, or experience knowledge expressing, or inferring and decision making, all of them are based on knowledge. Therefore, the research on knowledge representation propels information ages to change and develop from the elementary stage mainly with data processing to the high-grade stage mainly with knowledge processing. It has important influence in all of the fields of pattern recognition, natural language understanding, information processing, machine study, robot, automatic theorem proof, automatic programming, expert system, etc..Though now, there have been a lot of methods for knowledge representation, such as production, logic, semantic network, frame, script, yet it is not solved completely. Thus to explore new methods for knowledge representation is always one of important subjects in AI.ABSTRACTFortunately, conceptual structure, as a new method, expands the knowledge representation methods. It establishes the semantic explanation model for human being perception course, based on philosophy and psychology. Concept is an important component part of human thought, and is the thinking unit that reflects objective thing and its brief peculiar property. To form concept is such a procedure that supersedes between two directions along " from special to general " and " from general to special ". Concept structure is such a new method for knowledge representation, which is based on the course of cognition.Generally, natural language is the most direct method and symbol system most in use to express human ideas and pass information. Therefore natural language processing (NLP) research is the key for breakthrough progress in AI. Naturally, natural language describing and modeling is the base for the development of natural language understanding, and it decides the research process and the direction in the field of natural language understanding.Moreover, today, information on Internet is unimaginably growing. This requires intelligent information system not only to search for information automatically, but also to filter, refine, and translate information, based on the high level of understanding. The processing of these high level understandings must be and can only be based on semantics.Knowledge graph, a kind of special concept graph, is a new method to describe and process natural language. As a kind of representation for NLP, the method points out a new way for natural language describing and modeling and makes a big step forward to the semantic understanding of" know it and know why ".In this paper we study the concept structural theory and applications, especially the knowledge graphs for natural language processing. On the one hand, we propose the method to combine concept graphs with rules, and apply it in a development environment for expert system. On the other hand, we analyze the shortness of traditional NLP technology, and propose a new kind of NLP technology, which is suitable for Chinese language understanding. At the same time, knowledge graph, concept graph as well as other traditional knowledge representation methods are compared from different points of view.In Chapter 2 and Chapter 3, we study some special problems in concept graphs and its application in expert system. The approach to combine concept graphs with rules is used. The supplements between these two methods increase the validity of system.In Chapter 4, from the ontology view, knowledge graphs are studied in contrast with the other 3 kinds of important representations, and the originalness, generality and the validity of knowledge graph th...