Research On Intelligent Information Processing Models Based On Fuzzy Operators

Artificial intelligence is a hot research field in information science.The new generation of AI development plan upgrades AI technology to the national strategic level.The core of the new generation of AI is uncertain information processing.Intelligent information processing mainly faces uncertain information processing.The research of intelligent information processing model is the key of the new generation of artificial intelligence and intelligent information processing.Fuzzy reasoning and fuzzy neural network are important intelligent information processing models.Based on the fuzzy operators in fuzzy logic,this paper systematically studies the rules-based fuzzy reasoning and robustness,error of fuzzy reasoning,fuzzy associative memory and noise tolerance,and the application of fuzzy operators in image fusion.The main achievements and innovations are as follows:(1)The Lipschitz aggregation properties of fuzzy operators are studied.Triangular norm(t-norm)operators and implication operators satisfying Lipschitz conditions are found.For1-Lipschitz t-norm operators and implication operators,the properties of 1-k?-Lipschitz,copula and quasi-copula are studied.t-norm operators and implication operators with 1-k?-Lipschitz,copula and quasi-copula properties are found.(2)Aiming at the perturbation problem of rule-based fuzzy reasoning,a method of guaranteeing good robustness of fuzzy reasoning by choosing fuzzy operators is proposed.According to the properties of 1-Lipschitz,1-k?-Lipschitz,copula and quasi-copula of t-norm operators and implication operators,the effects of different types of Lipschitz t-norm operators and implication operators on the output of fuzzy reasoning are studied,and the effects of different types of Lipschitz t-norm operators and implication operators on the robustness of fuzzy reasoning under different perturbations are studied.When the t-norm operator and implication operator are both 1-k?-Lipschitz and quasi-copula,the fuzzy reasoning has good robustness,which is verified by experiments.(3)Aiming at the problem of rule-based fuzzy reasoning error,a control method of output error of fuzzy reasoning caused by the selection of fuzzy operators is proposed.According to the properties of 1-Lipschitz,1-k?-Lipschitz,copula and quasi-copula of t-norm operator and implication operator,the effects of different types of Lipschitz t-norm operator and implication operator on the output error of fuzzy reasoning are studied.The effects of different types of Lipschitz t-norm operator and implication operator on the output error of fuzzy reasoning under different perturbations are studied.When the t-norm operator and implication operator are both 1-k?-Lipschitz and quasi-copulas,the error of fuzzy reasoning caused by the selection of fuzzy operators can be effectively controlled,and the experimental verification is carried out.(4)Quasi-copula fuzzy associative memory model is proposed to solve the problem of noise tolerance in the implication fuzzy associative memory model.According to the properties of 1-Lipschitz,1-k?-Lipschitz,copula and quasi-copula of Lipschitz fuzzy operators,the effects of different types of Lipschitz t-norm operators and implication operators on the output of fuzzy associative memory are studied under different noise conditions.When the t-norm operator and implication operator are both quasi-copulas and 1-k?-Lipschitz in fuzzy associative memory,the fuzzy associative memory has good noise tolerance,which is verified by experiments.(5)A new method of image fusion using Lipschitz fuzzy operator is proposed.This method is simple,efficient and easy to implement in hardware.Experiments show that the fused image has good visual effect,clear details and clear targets.The experimental part of this paper is not only the verification of the proposed theory,but also the practical application of these theoretical results in intelligent information processing,such as face recognition,image processing,face Association and so on.