pith. sign in

Robustness via curvature regularization, and vice versa

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
abstract

State-of-the-art classifiers have been shown to be largely vulnerable to adversarial perturbations. One of the most effective strategies to improve robustness is adversarial training. In this paper, we investigate the effect of adversarial training on the geometry of the classification landscape and decision boundaries. We show in particular that adversarial training leads to a significant decrease in the curvature of the loss surface with respect to inputs, leading to a drastically more "linear" behaviour of the network. Using a locally quadratic approximation, we provide theoretical evidence on the existence of a strong relation between large robustness and small curvature. To further show the importance of reduced curvature for improving the robustness, we propose a new regularizer that directly minimizes curvature of the loss surface, and leads to adversarial robustness that is on par with adversarial training. Besides being a more efficient and principled alternative to adversarial training, the proposed regularizer confirms our claims on the importance of exhibiting quasi-linear behavior in the vicinity of data points in order to achieve robustness.

fields

cs.LG 1

years

2026 1

verdicts

UNVERDICTED 1

clear filters

representative citing papers

SORA: Free Second-Order Attacks in Fast Adversarial Training

cs.LG · 2026-05-30 · unverdicted · novelty 5.0

SORA is an adaptive step-size adversarial training algorithm that formalizes epsilon overfitting, introduces the PertAlign metric to predict catastrophic overfitting, and dynamically adjusts perturbations to achieve state-of-the-art robustness and clean accuracy with fixed hyperparameters.

citing papers explorer

Showing 1 of 1 citing paper after filters.

  • SORA: Free Second-Order Attacks in Fast Adversarial Training cs.LG · 2026-05-30 · unverdicted · none · ref 43 · internal anchor

    SORA is an adaptive step-size adversarial training algorithm that formalizes epsilon overfitting, introduces the PertAlign metric to predict catastrophic overfitting, and dynamically adjusts perturbations to achieve state-of-the-art robustness and clean accuracy with fixed hyperparameters.