An advanced deep learning framework for simulating information propagation dynamics

The warehouse model, based on differential equations, has been widely employed in the field of network information propagation for an extended period. Numerous studies have revolved around the construction, fitting and simulation of these models. However, there has not been a universal and efficient fitting method applicable to all warehouse models in the realm of information propagation, mainly due to the often challenging nature of solving differential equations in practical scenarios. In this article, we introduce a deep learning-based framework for simulating information propagation dynamics. This framework is grounded in a model that embeds a physical neural network and can be employed for fitting data from sentiment analysis platforms. We apply our framework to classic information propagation dynamic models, achieving favorable fitting results and consistent experimental outcomes, underscoring the advancement of our approach.