arXiv Machine Learning By Chenlong Wei, Hanling Wang, Chong Zhang, Qinkai Yu, Fangyu Wu, Xiaobo Jin

Structured Confidence Perturbation via Fisher Geometry for Adversarial Training

Read the original on arXiv Machine Learning →

The paper introduces two methods to improve adversarial training by targeting the model’s confidence structure. One method blends cross‑entropy with KL divergence to simultaneously degrade accuracy and distribution consistency, while the other uses Fisher geometry to generate perturbations that maximize distribution distance on the confidence manifold. Experiments show these techniques outperform existing approaches in robustness and transferability, especially for high‑confidence structural perturbations.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI
Sep 18

Information-Geometric Inverse Distillation for Enhancing Adversarial Transferability

The paper introduces Inverse Knowledge Distillation (IKD), an attack‑agnostic technique that enhances adversarial transferability by maximizing the discrepancy between benign and adversarial prediction distributions on a surrogate model. IKD employs a CE/KL‑equivalent soft‑label objective to push adversarial predictions away from a fixed benign anchor, leveraging Fisher‑sensitive surrogate directions. The authors provide theoretical analysis showing CE and KL induce identical gradients, derive a lower bound on Fisher‑subspace overlap, and demonstrate through extensive ImageNet experiments that IKD consistently improves black‑box attack performance across CNN, ViT, and defended models.

By Wenyuan Wu, Yuan Sun, Yingke Chen, Chao Su, Xi Peng, Dezhong Peng, Xu Wang
arXiv Machine Learning
1d ago

Representational Geometry as a Diagnostic for Adversarial Robustness

The paper investigates how adversarial training alters the internal representations of neural networks. Using Multilayer Perceptrons trained on MNIST under three different adversarial regimes, the authors embed training data with M-PHATE and analyze changes in geometry across epochs. They measure displacement, distributional distance, and nearest‑neighbor recovery on penultimate‑layer representations, finding that adversarial training induces structural changes that can rank defenses differently in terms of robust accuracy and representational stability.

By Brandon Ismalej, Alex Cloninger, Gal Mishne