Source detection in simplicial complexes based on dynamic message passing algorithm

The source detection problem, which aims to identify the seeds of false information propagating on social networks, has attracted extensive attention in the fields of network science and security. Previous information propagation models mainly focus on simple graphs (involving only pairwise interactions) and ignore the social reinforcement effect during the information propagation process. To address this, we build up a susceptible-infected-recovered (SIR) propagation model on simplicial complexes in this paper, which features the social reinforcement effect generated from higher-order interactions. Based on this model, we further propose a simplicial complex dynamic message passing (SCDMP) algorithm, and then develop an energy-based metric to infer the propagation source within simplicial complex networks. Experiments on real-world and synthetic networks demonstrate that, compared with baseline algorithms, the SCDMP algorithm exhibits significant advantages in detection accuracy, stability, and robustness.