This paper takes deep learning as the background of researcher design, combines the relevant cutting-edge research results in recent years, addresses the linguistic characteristics of Japanese and the problems faced by completing Japanese machine translation system, and determines the neural network structure of encoding-decoding for Japanese translation based on the characteristics of high similarity between Japanese and Chinese and after referring to the neural network architecture of English translation, and the basic structure and the corresponding improvement of the hidden layer unit calculation are carried out. The training model is optimized and an integrated Japanese machine translation system is implemented. Finally, the translation models of Japanese and Chinese intertranslation and Japanese and Chinese intertranslation are tested, and the optimal model fusion achieves a BLEU value of 39.52.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.