Abstract:
The tools are proposed to reduce ambiguousness of semantic models of statements, thus increasing tolerance of a man-machine dialogue and efficiency of knowledge acquisition systems. A semantic model of a situation (a case) is described by a plex grammar containing 4 sorts of symbols with various number of attaching points. A process of semantic text model synthesis is considered as a process of inference of a correct denotative structure adequate to human text perception. Operations of matching and joining of pattern graphs are proposed. Adequacy of a semantic model is assessed by checking its grammatical validity, matching elements of a new model to ones stored in a knowledge base, as well as detecting difference between properties, relations and links of new elements.