Game-theoretic approach to epidemic modeling of countermeasures against future malware evolution

Recently, vulnerability mining techniques that discover unknown vulnerabilities based on machine learning have been attracted much attention for protecting software. Although we benefit from these techniques for cyber security, they could be exploited by malicious attackers. For example, the literature has introduced a concept of future malware exploiting vulnerability mining techniques. It discovers vulnerabilities of hosts by performing vulnerability mining with the use of the computing resources of hosts infected with the malware. In this paper, we propose a game-theoretic approach to epidemic modeling for discussing how to counter such future malware evolution. In the proposed approach, we consider a countermeasure model that constructs a countermeasure group aiming to discover vulnerabilities earlier than malware or malicious attackers, and repair them to protect hosts not to get infected with the malware. This paper provides stochastic epidemic modeling for the countermeasure model, which represents the infection dynamics of future malware based on a continuous-time Markov chain under countermeasure environments. Furthermore, we apply evolutionary games on complex networks to the epidemic model in order to represent the selfish behavior of hosts participating in the countermeasure group. Through simulation experiments, we reveal strategies to efficiently counter the future malware evolution.

Automated summary

Recently, the use of machine learning for vulnerability mining has gained attention for software protection. However, these techniques could also be exploited by malicious attackers. This paper proposes a game-theoretic approach to epidemic modeling for discussing countermeasures against future malware evolution.