Intelligent Systems in Translation to Assist in Engineers' Training

The creation of machine translation technologies is currently an objective reality. Information technologies that already exist today that can be used to optimize translation are quite diverse. At this stage of computer-assisted translation (CAT) development, the issues of integrating artificial intelligence and translation activities (digital translation) and the use of neural machine translation (NMT) are relevant. Both technologies are quite effective, although they do not rule out the existence of certain difficulties in their use. So, for example, for a neural network, the difficulties are the translation of rare words, phraseological units, etc. Machine translation is a topic in which modern neural network algorithms have indeed achieved impressive successes. Thanks to advances in text generation, in the construction of vector representations of sentences that consider shades of meaning, as well as the use of the attention mechanism, modern machine translation tools often produce results that are almost indistinguishable from human ones. This paper describes the experiment carried out by the author to confirm the conclusions obtained from the results of his research.