Integrated modeling of language structures for linguistic processors of knowledge management and machine translation systems

The paper is dedicated to research and integrated modeling of cognitive linguistic representations of language structures and the mechanisms for resolution of syntactic ambiguity in the process of creating linguistic processors for intelligent knowledge processing and machine translation systems. The technique of constructing the hybrid cognitive linguistic model for representation of language structures and resolution of their ambiguity on the basis of logical linguistic rules and vector spaces has been developed and specified. The method presented is new and rests on the modern level of development in science and technology. The following tasks have been carried out: comparative research of classification methods has been made with respect to linguistic problems; the effective method of mapping the vector of natural language structures into the expanded space of attributes for classification of new language objects and structures has been worked out; the focal sample of parallel texts of business and scientific documents in Russian, English, and French has been developed; the expanded system of new categories enhancing the representational power of the initial grammar variant has been formed; the extended semantic networks were employed for unified representation of the matched language structures and the experiments of vector spaces method application for resolution of syntactic ambiguity of the key language structures were performed; the grammatical formalism and the algorithmic representation have been designed of the parser in which the difficulties of translation including the language transformations are taken into account.