Towards Robust Monitoring of Stealthy Diffusion

TL;DR

This paper introduces a game-theoretic framework for monitoring and defending networks against stealthy diffusion attacks, focusing on detection and containment under adversarial conditions.

Contribution

It proposes the (,)-Monitoring game model, analyzing attacker and defender strategies for robust network security against stealthy diffusion.

Findings

The (,)-Monitoring game effectively models stealthy diffusion attacks.

Robust defense strategies can be derived to minimize attack success in worst-case scenarios.

The approach outperforms stochastic guarantees in adversarial settings.

Abstract

In this work, we introduce and study the \emph{$(\alpha, \beta)$-Monitoring} game on networks. Our game is composed of two parties an attacker and a defender. The attacker can launch an attack by distributing a limited number of seeds (i.e., virus) to the network, and/or manipulate the propagation probabilities on a limited number of edges. Under our $(\alpha, \beta)$-Monitoring game, we say an attack is successful if and only if the following two conditions are satisfied: (1) the outbreak/propagation reaches $\alpha$ individuals, and (2) it has not been detected before reaching $\beta$ individuals. On the other end, the defender's ultimate goal is to deploy a set of monitors in the network that can minimize attacker's success ratio in the worst-case. Our work is built upon recent work in security games, compared with stochastic guarantees, our adversarial setting leads to more robust solutions in practice.