Associative semantics of situations and stories

The paper develops the thesis of the associativity of text perception, which sets connections between statements' elements on the basis of similarity of their attributes, subjects and relations or by means of the identification of a homomorphism between graphs. For this purpose we use a process-oriented domain-specific ontology and semantic models of situations, which are close to BPMN diagrams for business process management that are widely used in information systems. A situation graph matching algorithm is proposed, which is based on the method of a beam graph of a neighborhood of a specific radius. The semantic model of a text is composed through the substitution of fragments of the semantic models of precedents, not contradicting in terms of membership to a subsumption tree’s classes.
Texts of logical problems for primary and middle school students were used in the experiment. Results obtained by a problem solver, earlier developed as a programming training system, confirm that the machine correctly interprets text.