Fair adaptive influence maximization in social networks

Publication Type

Conference Proceeding Article

Publication Date

6-2026

Abstract

How can we influence different groups across a social network, reducing disparities among them while taking into consideration previous dissemination? The problem of Fair Influence Maximization (FairIM) has been primarily studied in non-adaptive settings, where influencers are chosen all at once without observing any information diffusion. In this paper, we study the FairIM problem in an adaptive setting (FairAIM), whereby seed nodes are chosen in multiple iterative batches, allowing the observation of influence spread to inform decisions in subsequent rounds, under the notion of maximin equity to maximize the minimum proportion of activated members among all groups. We propose two algorithmic frameworks for FairAIM: the first extends the greedy approach to adaptively select seeds to maximize the influence of the least well-off group; the second relaxes the discrete objective to a continuous space and uses a parameter β≥0 to balance the fairness and welfare objectives. The experiments evince that our algorithms significantly improve fairness while maintaining high overall welfare compared to prior IM methods.

Keywords

Adaptive influence maximization, Social networks, Algorithmic fairness

Discipline

Artificial Intelligence and Robotics | Databases and Information Systems

Research Areas

Intelligent Systems and Optimization

Publication

Proceedings of the 30th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2026, Hong Kong, China, June 9-12

First Page

296

Last Page

308

ISBN

9789819219254

Identifier

10.1007/978-981-92-1926-1_24

Publisher

Springer

City or Country

Cham

Additional URL

https://doi.org/10.1007/978-981-92-1926-1_24

This document is currently not available here.

Share

COinS