Overview of deep learning methods for neural machine translation

Chzhun Zhujyuj

The article discusses the main methods of neurolinguistic training. Three stages of machine translation development are revealed: rule-based, statistical, neural learning. Two popular methods of neural machine translation are described: recurrent and convolutional. Advanced neural machine learning methods are characterized: ConvS2S, Transformer, RNMT +. A forecast was made for the further development of neural text translation systems.

Keywords:machine translation, neural learning, neural learning methods, neurolinguistics, ConvS2S, Transformer, RNMT +.