WITHDRAWN: Fair Influence Maximization on Social Networks with Community Structure

• Activity Number: 319 - SLDS CSpeed 6
• Type: Contributed
• Date/Time: Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Octavio César Mesner and Ji Zhu and Liza Levina
• Companies: University of Michigan and University of Michigan and University of Michigan
• Keywords: social networks; fairness; influence maximization; stochastic block model; community structure; entropy maximization

Abstract

The use of social networks, virtual and in-person, to spread information is ubiquitous. Influence maximization (IM) algorithms use social networks to select individuals who will generate the greatest spread if seeded with information. However, in social networks with community structure, most IM algorithms may yield significant disparities in information coverage between communities, which could be problematic in settings such as public health messaging. While there are some algorithms to remedy this disparity using known individual attributes, such as race, none use the empirical community structure within the network itself. This work builds on existing algorithms for estimating community structure and then models information spread within and between communities. With this model, we determine optimal seed allocations balancing total maximum information coverage and expected fair coverage using maximum entropy.