The paper discusses the general issues of automated search of artifacts in rule-based knowledge bases (KB) based on logic with vector semantics in the VTF-logic variant. Cases of falsity of antecedent of rules at all admissible values of truth of input premises, existence of terms which are not used anywhere and generation of uncertain values of truth, emergence of contradictions are considered. Automation is considered as the organization of the direct attached logical inference opening artifacts of KB. The first two cases are identified by counting the number of each rule triggering and identifying terms that are not tied to the rules. The contradiction is revealed by the verification of the truth of the conclusions-hypotheses. The presence of a conclusion with truth (1; 1) (complete contradiction) signals a contradiction at one of the stages of reasoning, which is set by the back trace of the logical chain. A necessary stage of inference is to combine the evidence using 11-composition (the second form of disjunction). The paper also presents the principle of calculating the truth of the conclusion based on the truth of the premises, the strategy of combining evidence, numerical measures that can be used in the conclusion.
The paper is devoted to the problem of expert systems knowledge bases verification. The methodological basis of verification is logics with vector semantics in the VTF-logics form. The knowledge model is a rule-based system. The issues of algorithmization of contradictions and other problems detection are considered in the paper. Algorithmization is based on the characteristic features of in VTF-logics inference. It is shown that in the formalism under consideration, the anomalous truth of a small premise generates the same conclusion (a large premise is considered strictly true). Anomalies such as falsity, uncertainty, and contradiction are considered. The problems of reducing the computational complexity of algorithms are considered.
Работа посвящена обсуждению вопросов применимости неклассических логических исчислений к задаче верификации продукционных баз знаний. Рассмотрены возможности некоторых трёхзначных, четырёхзначных, а также нечётких логик. Показано, что хорошим подходом к верификации является использование логик с векторными семантиками в форме VTF-логик. Основанные на них экспер тные системы смогут верифицировать свои БЗ без привлечения дополнительных (и внешних по отношению к ЭС) архитектурных элементов.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.