Computer-aided Translation Algorithm Based on Artificial Intelligence and Deep Learning

Technologies such as DL (Deep learning), machine learning and neural network in the field of AI (Artificial Intelligence) have promoted the progress of CAT (Computer-Aided Translation). The neural network algorithm in DL algorithm can extract the language vector features of CAT translation with high accuracy in the form of learning and training bilingual words in the translation automatically. This paper studies the CAT algorithm based on AI and DL. In this paper, the model takes the model parameters trained by Chinese and English data with rich corpus as the initial values of Chinese and English CAT models by transferring the pivot parameters, which reduces the influence caused by the differences between Chinese and English in parameter transfer and improves the performance of Chinese and English CAT models with low resources. In addition, the model also combines the Tree-LSTM (Tree-Long Short-Term Memory) with the Bi-LSTM model to help the encoder extract effective information from the decoder. The research results show that the METEOR score of the model proposed in this paper is higher than that of the contrast model, which indicates that learning the tree-like structure of the source sentence through Tree-LSTM and using it as the context information of the decoder will help to enhance the connection between the long distance words understood by the model during decoding and improve the translation effect of the model.