Construction of a Neural Machine Translation Model based on Cloud LM Algorithm
Ying Du · 2024
In the context of global integration, the demand for machine translation is also increasing. Traditional machine translation methods are subject to certain limitations in accuracy and fluency. In response to this issue, this article intends to use the Cloud Language Model (Cloud LM), based on deep learning, to study the NMT (Neural Machine Translation) modeling method based on deep learning, and conduct a large number of experiments. In the robustness test, when there is no noise, the noise intensity is 0%, the BLEU (Bilingual Evaluation Understudy) score is 27.5, and the translation accuracy is 90%; When subjected to random noise, the noise intensity is 5%, the BLEU score is 26.2, and the translation accuracy is 88%. The experimental results show that the quality of multilingual translations based on Cloud LM is significantly better than traditional NMT models. The research results of this article can play a certain promoting role in the development of machine translation theory, and also have reference significance for other natural language processing problems.