The Historical Requirements and Epochal Characteristics Analysis of the Chinese Periodization of Marxist Ethnic Theory Integrating Multimodal Information

In order to continuously explore the new realm of Chineseization and modernization of Marxism’s ethnic theory, this paper proposes the study of Marxism’s ethnic theory based on multimodal information. Using LXMERT to initially encode the multimodal corpus of Marxist Ethnic Theory, the encoded multimodal features of Marxist Ethnic Theory are inputted into the noise filtering layer for training, which can more accurately determine the relationship between images and texts. In order to fuse the information of text and image in Marxist national theory multimodal, the multimodal attention mechanism is used to fuse the text and image context vectors, and the quality and efficiency of Marxist national theory multimodal features are improved by decoding. The specific types of manifestations of Marxist national theory that combine multimodal information are analyzed, and the research on the Chineseization of Marxist theory is also examined. According to the data, Marxist national theory has a multimodal sentiment recognition rate of 81.6%. Compared to the better performance of unimodal emotion recognition. The correct rate of students’ cognition of the era characteristics of Marxist national theory reaches 69.27%, which indicates that the students’ knowledge of the era characteristics of the Chineseization of Marxist national theory is mainly considered to be in general agreement and has not yet reached a perfect state.