Optimization of machine translation algorithm for English long sentences based on deep learning

Traditional artificial English long sentence translation efficiency is low, and the difference of the translator’s personal level will lead to uneven translation quality. The existing MT can not effectively solve the problems of cross-type ambiguity in English long sentence translation rules. This paper adopts the encoder-decoder structure. In the encoding stage, each word in the Chinese sentence is first mapped to a fixed-length word vector, and all the information of the entire sentence is compressed through the recurrent neural network. The attention model is introduced in the decoding process, so that the decoder pays more attention to the context-dependent words of the current translated word, and selects the translated word with the highest probability to generate the target sentence each time. Compared with traditional English long sentence translation methods, it has higher timeliness, higher accuracy, and stronger translation standards.