Research on Information Fusion Method of English Teaching Reform Based on Reinforcement Learning

Abstract

The information of English teaching reform has such attributes as different time and space, heterogeneous form, etc., which brings difficulties to multi-source data fusion and decision-making. The existing multimodal models may have some missing modal data or conflicts between modalities, which will bring difficulties to information fusion in English teaching reform. Therefore, an information fusion method of English teaching reform based on reinforcement learning is designed. The clustering algorithm based on multi-agent Reinforcement learning is adopted to implement the information mining of English teaching reform. Implement data cleaning for mining information, including missing data filling and duplicate data cleaning. The dynamic invariant special representation fusion network is designed, which is composed of the improved joint embedded domain separation network and the fusion representation to realize the information fusion of English teaching reform. The test results show that the average absolute error of the design method in fusion is lower than 0.700% as a whole, and the Pearson correlation coefficient is higher than 0.4 as a whole.