Lingo-Stylistic Analysis of Statistical and Neural Machine Translation

The urgency of the problem under investigation is caused by the increasing popularity of machine translation for the solution of various kinds of communicative tasks. The purpose of the article is to compare statistical and neural machine translation systems. The leading approach to the study of this problem was the linguo-stylistic analysis of linguistic material using software from Microsoft Translator. The main results of the article consist in a comparative analysis of the translation of simple and complex texts through statistical and neural machine translation systems, which led to the conclusion that the greatest number of errors is associated with the translation of semantic constructions. Materials of the article can be useful to the experts working in the field of machine translation, to students and all who are connected with computer linguistics.