Font Size: a A A

Research On Knowledge Fusion In AORBCO Model

Posted on:2024-05-02Degree:MasterType:Thesis
Country:ChinaCandidate:L S XiaoFull Text:PDF
GTID:2568307061969299Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Knowledge is the understanding and description of "consciousness",or knowledge is intelligence.However,the mainstream cognition of knowledge is that knowledge is the result of human understanding from the objective world in practice,there is no discussion of the subject of awareness.To establish a consensus-based knowledge to represent intelligence,expert systems,neural networks,knowledge graphs and other models have emerged in the field of artificial intelligence,these models do not take a deep research into the knowledge and will not be updated with the changes in the world.In order to build a comprehensive knowledge base that is self-awareness and can be updated with the world,this thesis divides the knowledge fusion into four parts: descriptive,process,emotional and strategic knowledge fusion on the basis of knowledge representation research,which makes the Ego in the AORBCO model can learn new knowledge and form its own understanding,and is the key for the model to evolution and problem planning.The following are the main research of the study.1)According to the essence of knowledge,AORBCO model is chosen as the method of knowledge representation.To address the problem of existing knowledge only defines intelligence in terms of sources,determinants and components,and separates intelligence from knowledge,this paper redefines knowledge based on self-consciousness as the understanding and description of "consciousness".And existing knowledge representations are analyzed and compared from three new dimensions: the four characteristics of knowledge,the role of knowledge and the composition of knowledge.It was found that these knowledge can only represent a part of the intelligence.The AORBCO model is chosen for knowledge representation because the knowledge representation method in the AORBCO model can solve the problems of existing knowledge representation and can represent intelligence more comprehensively.2)The four kinds of knowledge are divided according to their different roles and their fusion methods are studied.The knowledge fusion in the AORBCO model is to map the things perceived in the real world to the abstract world,to form the recognition of the new world and to complete the knowledge update.Descriptive knowledge is a kind of knowledge that describes the state of things,in the study of descriptive knowledge fusion,a CANE-EA algorithm is proposed for the entity alignment problem,which aligns entities according to the class that entity belong to and the surrounding entities,so that Ego establishes the connection from the real world to the belief in the abstract world.Process knowledge is the operation of changing the state of things,in the process knowledge fusion research,the parameters and parameter constraints are defined first,and a PARO-FA algorithm based on parameters,precondition and postcondition is proposed to determine whether the function is same,so that the Ego can learn function.Strategic knowledge is the selection of process knowledge,in the strategic knowledge fusion study,Strategy Knowledge Graph(SKG)is used to represent strategic knowledge,so that Ego can reason when planning a problem.Affective knowledge is the knowledge of value judgment,in the affective knowledge fusion study,the changing relationship of weights in descriptive and strategic knowledge is given so that Ego forms value judgments about things and makes changes.3)Three experiments were designed in the platform of the AORBCO model to verify the feasibility of the fusion algorithm and applied in intelligent Web to verify the availability of the fusion algorithm.In the experiment of descriptive knowledge fusion,the feasibility of entity fusion algorithm is verified by fusing several triads of knowledge.In the process and strategic knowledge fusion experiments,the feasibility of the process knowledge fusion algorithm and the strategic knowledge fusion algorithm are verified by fusing new functions.The results show that these proposed algorithm allows Ego’s belief,function,and SKG to be updated,and providing a way for Ego to learn new knowledge and form a new understanding of the world.
Keywords/Search Tags:AORBCO model, knowledge, knowledge representation, knowledge fusion, entity alignment, strategic knowledge graph
PDF Full Text Request
Related items