Research on Translation Strategies of International Publicity Integrating Artificial Intelligence Translation Technology

In the stage of globalization, the translation of international publicity, as an important medium of language and culture communication, has aroused widespread concern and discussion in society. In this paper, a machine translation (MT) semantic model based on feature extraction algorithm is proposed, through which the relationship between fuzzy semantics is mapped to better extract the features of sentence content. In this model, a template matching model for automatic calibration of MT is constructed by using differentiated semantic modification method, and a semantic tree for automatic calibration of MT is established by combining with syntax analysis method. The experimental results show that: compared with other algorithms, the error of this algorithm is lower, mean square error (MSE) is about 4.15, mean absolute error (MAE) is about 1.21; At the same time, the accuracy of this algorithm is higher than the other two algorithms, which can reach 95.97%, and it has certain superior performance. The comparative experimental results show that the designed system can get the optimal translation and the translation accuracy is high.