aiqbench analysis
This research introduces Calibrated Adversarial Sampling (CAS), a method that reformulates multi-attack adversarial training as a multi-armed bandit problem to reduce computational overhead. By dynamically balancing exploration and exploitation, CAS mitigates parameter drift while improving robustness against diverse perturbations without the prohibitive costs of existing approaches. The work offers a scalable framework for enhancing deep neural network security through efficient optimization techniques central to machine learning safety.
Stabilizing Multi-Attack Adversarial Training via Bandit Optimization
arXiv:2511.12265v2 Announce Type: replace Abstract: Deep Neural Networks (DNNs) remain vulnerable to diverse adversarial perturbations, motivating multi-attack adversarial training (AT) for improved robustness. However, existing methods either incur prohibitive overhead by computing all attacks at each iteration, or rely on stochastic sampling over adversarial examples, which may cause excessive parameter drift. To address these issues, we propose Calibrated Adversarial Sampling (CAS), an efficient and stable framework that reformulates multi-attack AT as a multi-armed bandit optimization problem. By sampling a single attack per iteration that dynamically balances exploration and exploitation, CAS significantly reduces training cost while mitigating optimization conflicts across attacks and controlling excessive parameter drifts. Extensive experiments demonstrate that CAS achieves superior overall robustness at low computational cost, offering a scalable and principled approach to robust generalization against multi-attack settings. Our code is available at https://github.com/1240148048/CAS.
Abstract:Deep Neural Networks (DNNs) remain vulnerable to diverse adversarial perturbations, motivating multi-attack adversarial training (AT) for improved robustness. However, existing methods either incur prohibitive overhead by computing all attacks at each iteration, or rely on stochastic sampling over adversarial examples, which may cause excessive parameter drift. To address these issues, we propose Calibrated Adversarial Sampling (CAS), an efficient and stable framework that reformulates multi-attack AT as a multi-armed bandit optimization problem. By sampling a single attack per iteration that dynamically balances exploration and exploitation, CAS significantly reduces training cost while mitigating optimization conflicts across attacks and controlling excessive parameter drifts. Extensive experiments demonstrate that CAS achieves superior overall robustness at low computational cost, offering a scalable and principled approach to robust generalization against multi-attack settings. Our code is available at this https URL.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV); Optimization and Control (math.OC)
arXiv:2511.12265 [cs.LG]
(or arXiv:2511.12265v2 [cs.LG] for this version)
Full text under the publisher's open licence · 161 words
Related coverage
[AINews] Reality Checks on AI News (Yegge shuts down Gas Town, Databricks’ +60% Astra cost)
Latent Space · Sep 17, 2026
Research Note: Filtering Subversion-Relevant Information From Pretraining Data Is Feasible
LessWrong · Oct 5, 2026
Angles of attack for continual learning safety
LessWrong · Jun 16, 2026
The Alignment Problem: Computer Science Meets Analytic Philosophy
AI Alignment Forum · Oct 18, 2026