Support Vector Machines Under Adversarial Label Noise

In adversarial classification tasks like spam filtering and intrusion detection, malicious adversaries may manipulate data to thwart the outcome of an automatic analysis. Thus, besides achieving good classification performances, machine learning algorithms have to be robust against adversarial data manipulation to successfully operate in these tasks. While support vector machines (SVMs) have shown to be a very successful approach in classification problems, their effectiveness in adversarial classification tasks has not been extensively investigated yet. In this paper we present a preliminary investigation of the robustness of SVMs against adversarial data manipulation. In particular, we assume that the adversary has control over some training data, and aims to subvert the SVM learning process. Within this assumption, we show that this is indeed possible, and propose a strategy to improve the robustness of SVMs to training data manipulation based on a simple kernel matrix correction.