Adversarial attacks on voter model dynamics in complex networks

TL;DR

This paper demonstrates that small, strategic perturbations in complex social networks can significantly distort voter model dynamics, potentially reversing majority opinions without detection.

Contribution

It introduces a simple adversarial attack method that effectively manipulates voter model outcomes in complex networks, highlighting vulnerabilities in opinion dynamics.

Findings

Adversarial attacks can invert majority opinions in social networks.

Effectiveness increases with network size and density.

Small perturbations are sufficient to distort opinion outcomes.

Abstract

This study investigates adversarial attacks conducted to distort voter model dynamics in complex networks. Specifically, a simple adversarial attack method is proposed to hold the state of opinions of an individual closer to the target state in the voter model dynamics. This indicates that even when one opinion is the majority, the vote outcome can be inverted (i.e., the outcome can lean toward the other opinion) by adding extremely small (hard-to-detect) perturbations strategically generated in social networks. Adversarial attacks are relatively more effective in complex (large and dense) networks. These results indicate that opinion dynamics can be unknowingly distorted.