English Long Sentence Machine Translation Algorithm Based on U-Net Generation Adversarial Network

U-Net Generative Adversarial Network (GAN) algorithm aims to solve the challenge of translating long sentences in machine translation, and improve the accuracy and fluency of such sentence translation. This paper further discusses the application of generative adversarial network (GAN) in machine translation, and then improves the translation quality through the adversarial training of the generation and discrimination network, and seeks to deal with the problem of translation accuracy skillfully through the in-depth analysis of the translation of long sentences. It can be seen from the experimental data that the translation speed of U-Net generated adversarial network’s English long sentence machine translation algorithm accounts for 92.25% of the data within 1.5 seconds, and the translation accuracy rate reaches 8~9 points. It can be seen that the organic combination of U-Net architecture and generative adversarial network provides a novel and effective solution to the problem of long sentence translation, and ultimately significantly improves the translation accuracy and deep meaning.