Competitive spreading dynamics under sentinel intervention in higher-order community networks.

Abstract

Misinformation spreading on social networks is jointly shaped by community structure, higher-order reinforcement, and intervention. We formulate a sentinel-triggered competitive susceptible-infected-susceptible-AB model on a simplicial complex with two equal-sized communities. Sentinels respond to local misinformation exposure by temporarily becoming truth sources, coupling local detection and corrective response with higher-order reinforcement and cross-community invasion. We derive a heterogeneous mean-field description and compare its predictions with Monte Carlo simulations. The results reveal extinction, community-localized persistence, and network-wide spreading, separated by discontinuous transitions and initial-condition-dependent hysteresis in the examined parameter ranges. Increasing the sentinel fraction or the pairwise truth-spreading strength suppresses the high-misinformation state, whereas a higher sentinel activation threshold delays the response. Intercommunity coupling promotes misinformation invasion under weak intervention but can support truth spreading when sentinel intervention is sufficiently strong. Extensions to unequal community sizes and more general community structures remain topics for future work.