Content area
Internet technology continues to develop, and the role of machine translation in the field of English language is becoming more and more prominent. The research is based on computer translation technology, which involves constructing a translation model and a syntax error correction model to enable multi-scenario English language applications. The BERT model, which has a good performance in natural language processing, is combined with the neural machine translation model, and through the internal fusion and dynamic weighting of multiple attention mechanisms to improve the model’s learning ability of the representation information, to realize the English translation model based on the improved Masking-BERT enhancement. The English grammar error correction task is considered a machine translation task, and Transformer is used as the skeleton model to build the English grammar error correction model. The English translation model in this paper shows good application results, with an accuracy rate of 72% to 87%, and has faster response practices and higher satisfaction evaluations. The English grammar error correction model is capable of detecting and correcting all kinds of grammatical errors, and the scores of each evaluation index are above 78%. The language model constructed in this paper has high effectiveness and feasibility, and can provide effective support and help for English translation and grammatical error correction.