PROBLEMS OF MACHINE TRANSLATION IN THE RUSSIAN–ENGLISH DIRECTION: A COMPARATIVE ANALYSIS OF MODERN TRANSLATION SYSTEMS

Abstract

This article examines Russian–English machine translation using Yandex, Google, DeepL, and ChatGPT and evaluates their performance based on BLEU, METEOR, and POS-BLEU metrics. The findings reveal differences in lexical, semantic, and syntactic accuracy, with Yandex demonstrating the strongest overall performance. The study also highlights the limitations of automatic translation evaluation metrics and emphasizes the importance of combining them with expert assessment and post-editing, particularly when translating scientific and specialized texts.