Robust Design of Machine Translation System Based on Convolutional Neural Network

Aiming at the problem of low load robustness coefficient and recovery robustness coefficient of machine translation system in different scenarios and working conditions, which leads to poor robustness, convolutional neural network algorithm is used to optimize the robustness of machine translation system. The operation process of the machine translation system is simulated through the steps of corpus preprocessing according to the composition and working principle of the machine translation system, word alignment processing, and phrase extraction. Obtain the load data of the machine translation system, and with the support of building a convolutional neural network model, according to the measurement results of the vulnerability of the machine translation system, use the convolutional neural network algorithm to determine the system load scheduling amount. The robust controller is selected as the executive element to complete the robust design of the machine translation system. The experimental results show that the machine translation system designed by the optimization method has higher load robustness coefficient and recovery robustness coefficient under different scenarios and operating conditions, which confirms that the robustness design effect of the optimized machine translation system is better.