Based on the theory of the quasi-truth degrees in two-valued predicate logic, some researches on approximate reasoning are studied in this paper. The relation of the pseudo-metric between first-order formulae and the ...Based on the theory of the quasi-truth degrees in two-valued predicate logic, some researches on approximate reasoning are studied in this paper. The relation of the pseudo-metric between first-order formulae and the quasi-truth degrees of first-order formulae is discussed, and it is proved that there is no isolated point in the logic metric space (F, ρ ). Thus the pseudo-metric between first-order formulae is well defined to develop the study about approximate reasoning in the logic metric space (F, ρ ). Then, three different types of approximate reasoning patterns are proposed, and their equivalence under some condition is proved. This work aims at filling in the blanks of approximate reasoning in quantitative predicate logic.展开更多
Truth maintenance systems become the very useful tools in artificial intelligence. Existing truth maintenance systems can’t deal with nonmonotonic reasoning effectively. They have limitations ...Truth maintenance systems become the very useful tools in artificial intelligence. Existing truth maintenance systems can’t deal with nonmonotonic reasoning effectively. They have limitations in the representation of nonmonotonic justifications. We present stratified truth maintenance systems which introduce priorities among justifications. The stratified truth maintenance systems can deal with nonmonotonic reasoning more effectively and can be applied in many useful areas.展开更多
Case Based Reasoning (CBR) is a powerful problem solving technique in AI, but the traditional CBR techniques have its limitations. We hybridized stratified ATMS and ANN for CBR which can deal with case representation...Case Based Reasoning (CBR) is a powerful problem solving technique in AI, but the traditional CBR techniques have its limitations. We hybridized stratified ATMS and ANN for CBR which can deal with case representation, case retrieving, case adapting, learning from failure more effectively. The structure of our CBR system and algorithms of case base reasoning in our CBR system were presented.展开更多
基金National Natural Science Foundation of China (No. 60875034)Spanish Ministry of Education and Science Fund,Spain (No.TIN-2009-0828)Spanish Regional Government (Junta de Andalucia) Fund,Spain (No. P08-TIC-3548)
文摘Based on the theory of the quasi-truth degrees in two-valued predicate logic, some researches on approximate reasoning are studied in this paper. The relation of the pseudo-metric between first-order formulae and the quasi-truth degrees of first-order formulae is discussed, and it is proved that there is no isolated point in the logic metric space (F, ρ ). Thus the pseudo-metric between first-order formulae is well defined to develop the study about approximate reasoning in the logic metric space (F, ρ ). Then, three different types of approximate reasoning patterns are proposed, and their equivalence under some condition is proved. This work aims at filling in the blanks of approximate reasoning in quantitative predicate logic.
文摘Truth maintenance systems become the very useful tools in artificial intelligence. Existing truth maintenance systems can’t deal with nonmonotonic reasoning effectively. They have limitations in the representation of nonmonotonic justifications. We present stratified truth maintenance systems which introduce priorities among justifications. The stratified truth maintenance systems can deal with nonmonotonic reasoning more effectively and can be applied in many useful areas.
文摘Case Based Reasoning (CBR) is a powerful problem solving technique in AI, but the traditional CBR techniques have its limitations. We hybridized stratified ATMS and ANN for CBR which can deal with case representation, case retrieving, case adapting, learning from failure more effectively. The structure of our CBR system and algorithms of case base reasoning in our CBR system were presented.
