摘要
人工智能以知识为基础,然而知识往往是可错的.当知识被发现是错误时,就要撤销该错误知识及依赖于该错误知识的知识.基于对Forget机制、TMS和ATMS的改进,提出一种新的知识表示方法,即带有依赖关系的知识库.基于命题逻辑程序,给出带有依赖关系的知识库的示例,同时还给出1个规则加入算法和2个规则撤销算法.第2个规则撤销算法可以处理循环依赖关系.带有依赖关系的知识库,是一种全新结构的知识库,在知识表示、信念修正和辩论推理中具有应用的前景.
Knowledge plays an important role in artificial intelligence,but there is the possibility that knowledge turns out to be erroneous.When a proposition turns out to be erroneous,this erroneous proposition and the other propositions,which depend on the erroneous proposition,will be retracted from the knowledge base.Based on TMS and ATMS,this paper presents a novel representation of knowledge,i.e.knowledge base with dependencies.This paper applies this idea to propositional logic program.Given a logic program,consequences together with their dependencies are recorded in the consequence set.Based on the consequences set with the dependencies,this paper proposes the algorithm of rule addition and two algorithms of rule deletion.The second algorithm of rule deletion can handle the problem of circular dependency.As a conclusion,this paper argues that justifications of knowledge should be stored in knowledge bases and different kinds of knowledge have different kinds of justifications.In this paper,this idea is applied to propositional logic program.In the future,we will extend this method to predicate logic and inductive logic.Based on knowledge base with dependencies,the retractions of knowledge become more efficiency.Therefore,this method provides a new foundation for belief revision and argument.
出处
《厦门大学学报（自然科学版）》
CAS
CSCD
北大核心
2011年第6期989-993,共5页
Journal of Xiamen University：Natural Science
关键词
知识库
依赖关系
真值维护系统
逻辑程序
knowledge base
dependency
truth maintenance system
logic program
