← Back to all news
arXiv Machine Learning By Yannick Lunk, Atell Yehor Krasnopolsky, Damien Garreau, Leon Bungert

Explaining the Saliency Map Sparsity of Adversarially-Trained Neural Networks

Read the original on arXiv Machine Learning →

The paper investigates why saliency maps of adversarially‑trained neural networks appear sparse. It provides a theoretical explanation for two‑layer ReLU networks, showing that as data size and neuron count grow, minimizers converge to a Bayes classifier with minimal gradient and Barron norm. The sparsity arises because α−−‑training with λ−1 attacks yields anisotropic gradient norms that favor axis‑aligned, sparse gradients, a claim supported by experimental evaluation of gradient α‑norms and sparsity thresholds.

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.

Related stories

arXiv Machine Learning
1d ago

Estimating Implicit Regularization in Deep Learning

arXiv:2605.05436v2 Announce Type: replace-cross Abstract: Deep learning systems are known to exhibit implicit regularization (alt. implicit bias), favoring simple solutions instead of merely minimizi...

By Joseph H. Rudoler, Kevin Tan, Giles Hooker, Konrad P. Kording
safety
More like this →
arXiv AI
Jun 4

Efficient Adversarial Attacks on High-dimensional Offline Bandits

arXiv:2602. 01658v2 Announce Type: replace-cross Abstract: Bandit algorithms have recently emerged as a powerful tool for evaluating machine learning models, including generative image models and large language models, by efficiently identifying top-performing candidates without exhaustive comparisons.

By Seyed Mohammad Hadi Hosseini, Amir Najafi, Mahdieh Soleymani Baghshah
llmsreinforcement-learningsafety
More like this →
arXiv AI
Sep 18

Exploring Sparsity and Smoothness of Arbitrary Lp Norms in Adversarial Attacks

The paper investigates how the choice of the π parameter in λπ norm-constrained adversarial attacks influences the sparsity and smoothness of the perturbations. By applying two established sparsity metrics and introducing three new smoothness measures—including one based on first-order Taylor approximations—the authors perform extensive experiments on real-world image datasets and various neural network architectures. Their results indicate that λρ norms with π values between 1.3 and 1.5 consistently provide the best balance between sparsity and smoothness, challenging the common use of λ1 or λ2 norms.

By Christof Duhme, Florian Eilers, Xiaoyi Jiang
llmsefficiencysafety
More like this →
arXiv AI
6d ago

Sparse Fine-Tuning for Parameter-Efficient Adversarial Training

The paper introduces Sparse Adversarial Fine‑Tuning (SAFT), a method that identifies a small set of Robustness‑Critical Parameters (RCPs) through adversarial gradient saliency and updates only these parameters during adversarial training. By focusing on this sparse subspace, SAFT achieves higher clean and robust accuracy than LoRA‑based parameter‑efficient methods and matches or surpasses full‑parameter adversarial training while training only about 5% of the parameters. The approach is validated across multiple architectures and datasets, demonstrating its effectiveness in reducing memory cost without sacrificing robustness.

By Zhaoxin Wang, Zihang Ding, Handing Wang
fine-tuningsafety
More like this →