Security Evaluation of Support Vector Machines in Adversarial Environments
This work addresses security vulnerabilities in SVMs for applications like malware detection, which is crucial for deploying robust machine-learning systems, though it is incremental as it builds on existing adversarial machine-learning concepts.
The paper tackles the problem of evaluating the security of Support Vector Machines (SVMs) in adversarial environments by introducing a formal framework and demonstrating the feasibility of evasion, poisoning, and privacy attacks in real-world security applications, with experiments made reproducible through open-source code and datasets.
Support Vector Machines (SVMs) are among the most popular classification techniques adopted in security applications like malware detection, intrusion detection, and spam filtering. However, if SVMs are to be incorporated in real-world security systems, they must be able to cope with attack patterns that can either mislead the learning algorithm (poisoning), evade detection (evasion), or gain information about their internal parameters (privacy breaches). The main contributions of this chapter are twofold. First, we introduce a formal general framework for the empirical evaluation of the security of machine-learning systems. Second, according to our framework, we demonstrate the feasibility of evasion, poisoning and privacy attacks against SVMs in real-world security problems. For each attack technique, we evaluate its impact and discuss whether (and how) it can be countered through an adversary-aware design of SVMs. Our experiments are easily reproducible thanks to open-source code that we have made available, together with all the employed datasets, on a public repository.