FIMH algorithm ensures fair influence spread in hypergraphs

Influence maximization—selecting initial 'seed' nodes to maximize the spread of information, products, or behaviors—is a cornerstone of viral marketing and public health campaigns. However, in networks with pronounced community structure, traditional methods often concentrate influence in dominant groups, leaving minority communities underserved. While fair influence maximization (FIM) has been studied for pairwise networks, real-world interactions like group chats, co-authorship teams, or event attendance are better modeled as hypergraphs (where edges connect more than two nodes).

To address this gap, researchers Zoë Abhelakh, Tianrui Mao, and Huijuan Wang propose FIMH (Fair Influence Maximization in Hypergraphs), a heuristic seed-selection algorithm that operates under the Susceptible-Infected Contact Process (SICP) model. FIMH iteratively estimates each candidate node's contribution to both total influence and fairness, then selects the seed that best balances these objectives using a parameter-free utopia-distance criterion. Experiments on seven real-world hypergraphs—ranging from social networks to collaboration graphs—show that FIMH matches state-of-the-art influence levels while cutting influence disparity. The work provides a practical tool for campaigns that need equitable reach across diverse groups.