Optimization of Bayesian Inference Algorithm in Multimodal Data Fusion
Jiefeng Deng, Wenming Wang · 2024
This paper proposes an optimized Bayesian inference algorithm, which aims to improve fusion accuracy and computational efficiency by improving the model structure and introducing an adaptive weighting mechanism. Specifically, this paper optimizes the algorithm’s computational process by adjusting the structure of the Bayesian network, using methods such as variational inference and distributed computing, and reduces the computational burden in the data fusion process. In addition, this paper also proposes a new data uncertainty modeling strategy that can better handle the noise and missing value problems of different modal data. Through experimental verification in multiple scenarios such as smart home, medical diagnosis, and autonomous driving, the results show that the optimized algorithm performs significantly better than the traditional Bayesian inference method on multiple data sets. In experiment A, the fusion accuracy of the optimized algorithm is improved by 10%, and the computational time is reduced by 15%; in experiment B, the error rate is reduced by 12%, and the computational efficiency is improved by 18%. The experimental results show that the proposed optimization algorithm can effectively improve the accuracy and efficiency of multimodal data fusion, and has good scalability and real-time performance.