Exploring the Hyperparameter Landscape of Adversarial Robustness
This work addresses the practical challenge of tuning adversarial training for machine learning practitioners, but it is incremental as it builds on existing methods without introducing new paradigms.
The paper investigates how hyperparameter settings affect adversarial training, finding that robustness depends on a few key parameters and that no universal settings exist, and proposes a hyperparameter optimization method to maximize robustness while limiting accuracy loss.
Adversarial training shows promise as an approach for training models that are robust towards adversarial perturbation. In this paper, we explore some of the practical challenges of adversarial training. We present a sensitivity analysis that illustrates that the effectiveness of adversarial training hinges on the settings of a few salient hyperparameters. We show that the robustness surface that emerges across these salient parameters can be surprisingly complex and that therefore no effective one-size-fits-all parameter settings exist. We then demonstrate that we can use the same salient hyperparameters as tuning knob to navigate the tension that can arise between robustness and accuracy. Based on these findings, we present a practical approach that leverages hyperparameter optimization techniques for tuning adversarial training to maximize robustness while keeping the loss in accuracy within a defined budget.