Non-monotonic Inference Modelling Of Knowledge Base With Subjective Conjecture Based On Satisfaction System

Based on the classical logic model of traditional knowledge base,this paper constructs a non-classical and non-monotonous logic model by introducing hypothesis,conjecture and some practical reasoning examples.The model has the characteristics of knowledge Increase or ignorance-preserving.It obviously improves the scope of knowledge base utilization and provides a logical basis for knowledge innovation.The proposed models are preference,rationality and abduction,respectively.This is a new and global knowledge base construction and application mechanism.This mechanism is based on the philosophical consideration of the evolutionary process of intelligence.From low to high,it is divided into four dimensions: time,consciousness,ethics and intelligence.Time is the most basic dimension,based on the present,there are many possibilities in the future,namely time branch.This branch reflects the uncertainty and possibilities of the future,thus introducing the diversity of consciousness,that is,the branch of consciousness.In fact,the branch of consciousness implies the diversity of human thought and the reasoning process of knowledge.This reasoning process is a process of choosing the future.This process of choice is generally a process of reasoning which is beneficial to human beings themselves,thus resulting in ethical branch.In ethical branch,artificial intelligence is generated,and at the same time,whether artificial intelligence can surpass human intelligence is questioned,which results in intelligent branch.Artificial intelligence and super-artificial intelligence are proposed in intelligent branch.Formalize the evolutionary process of intelligence in philosophy,then it is cumulative logic,preference logic and rational logic.In the course of the evolution of intelligence,with the process of human reflection,it can be formalized as a process of abductive reasoning.Thus,according to the evolution of AI,a new application and mechanism of knowledge set is proposed.Based on the classical knowledge set,hypothesis,conjecture and example are reduced to default knowledge set,and three kinds of default knowledge set modeling process and corresponding reasoning model are proposed.In the process of constructing the preference reasoning model,the formulas of knowledge set are assigned and the corresponding limited world set is established.Secondly,the standard model of preference relation and preference reasoning relation between the world is established.Finally,the determinants of the basic default knowledge set and the default knowledge set are constructed according to the established preference standard model,in which the basic default set and the determinant are equivalent.On the basis of the preference standard model,the ranks of each world in the world concentration are defined,and thefeature set of reasonable reasoning relationship is constructed.The feature set of rational reasoning is equivalent to the relationship of rational reasoning.The rank of the world is established,and the model of reasonable reasoning relationship is established accordingly.To define the construction process of the abductive reasoning model under the interpretation relationship.By using the cutting method defined in the semantic model of knowledge set,an ordering relation of the world with inclusion relation is established.This order relation has the property of differentiable continuity.Then the corresponding contraction and expansion operators are defined under the order relation.Contraction and expansion operators can be expressed as corrosion and expansion operators in knowledge dynamics equivalently.Therefore,the corresponding conjecture generation process is defined by the abductive reasoning model.Through the above three modeling methods,the reasoning on knowledge set can have the characteristics of knowledge increase or ignorance-preserving.