SPECIFIC FEATURES OF IDENTIFYING SPECIAL CHARACTERS AND TERMS BY MACHINE TRANSLATION SYSTEMS

The purpose of the article is to investigate the ability of machine translation systems to recognize special characters and terms based on the results of translating from English to Ukrainian and vice versa to identify problems and establish the causes of their occurrence. The object of the research is the results of identifying spelling and terminology features in machine translation of scientific texts in the field of sociolinguistics on such topics as bilingualism and language conflicts translated from English to Ukrainian and vice versa, using three machine translation systems; the subject is special characters and terms identified by machine translation systems as the result of translating from English to Ukrainian and vice versa, to identify problems, relevant and irrelevant variants. The research in the scope of translation theory and computational linguistics uses both general scientific methods and purely linguistic methods such as: analysis, synthesis, comparative and correlative methods. The results of the work are to establish identification problems that occur in machine translation of scientific texts and their analysis. Scientific texts may contain special characters to represent specific features of phonetics or spelling of a foreign language word or term and general terms may have spelling differences in English or Ukrainian. Special characters such as quotation marks, dashes and transliteration symbols may cause difficulties identifying them by systems. The obtained results of the study allow to conclude. The practical value is in the possibility of using the results in research in the field of computational linguistics, linguistic modelling and corpus linguistics, translation theory and practice, practical course of translation. The conclusions of the research can be seen in the further detailed explanation of the identified problems of special characters and terms, to propose, on the basis of that, ways to optimize machine translation.