Im-balanced: Influence maximization under balance constraints

Influence Maximization (IM) is the problem of finding a set of influential users in a social network, so that their aggregated influence is maximized. IM has natural applications in viral marketing and has been the focus of extensive recent research. One critical problem, however, is that while existing IM algorithms serve the goal of reaching a large audience, they may obliviously focus on certain well-connected populations, at the expense of key demographics, creating an undesirable imbalance, an illustration of a broad phenomenon referred to as algorithmic discrimination. Indeed, we demonstrate an inherent trade-off between two objectives: (1) maximizing the overall influence and (2) maximizing influence over a predefined protected" demographic, with the optimal balance between the two being open to different interpretations. To this end, we present IM-Balanced, a system enabling end users to declaratively specify the desired trade-off between these objectives w.r.t. an emphasized population. IM-Balanced provides theoretical guarantees for the proximity to the optimal solution in terms of both objectives and ensures an efficient, scalable computation via careful adaptation of existing state-of-the-art IM algorithms. Our demonstration illustrates the effectiveness of our approach through real-life viral marketing scenarios in an academic social network.