analiza aiqbench
This research introduces GRAPE, a novel training framework that enhances adversarial robustness by progressively exposing model parameters rather than optimizing them simultaneously. By guiding the evolution of parameter space toward high-pressure modules, the method achieves significant accuracy gains and reduced computational costs compared to standard fixed-structure approaches. The findings highlight how the order of optimization influences final solution quality, offering a more efficient path for building secure neural networks under strict resource constraints.
GRAPE: Guided Parameter-Space Evolution for Compact Adversarial Robustness
arXiv:2606.14865v1 Announce Type: new Abstract: Adversarial Training (AT) improves neural network robustness, but most methods train a fixed parameter space from the start. This paper asks whether the order in which parameters become optimizable can affect the final robust solution, even when the final architecture or computation budget is controlled. We propose GRAPE, Guided Parameter-Space Evolution, a training framework for compact adversarial robustness. GRAPE combines parameter-space stabilization with progressive hidden expansion: it stabilizes robust optimization in the currently exposed space, gradually releases new optimizable dimensions, and uses an adversarial spectral utilization score to guide newly released capacity toward high-pressure modules. In contrast to fixed-structure AT, GRAPE treats robust model learning as a process of progressive parameter-space exposure and evolution. Under the standard $\ell_\infty$ threat model on CIFAR-10, with fixed-structure ResNet-18 AT as a controlled reference, GRAPE improves PGD-20 robust accuracy from 51.70% to 56.94% at a nearly matched computation budget with a FLOPs ratio of 1.009x, while reducing parameter count by about 21.4%. A sequential grow variant with the same final ResNet-18 architecture reaches 56.52% PGD-20 robust accuracy, indicating that the gain is not only due to final architecture differences but also to the parameter-space exposure path. These results suggest that guided parameter-space evolution can yield compact and robust parameter configurations under matched computation.
Authors:Zhiyuan Ye (1), Xiangyu Zhou (2), Ji Qi (2), Hao Zhang (1), Yi Zhou (2) ((1) University of Science and Technology of China, (2) China Mobile (Suzhou) Software Technology Co., Ltd.)
Abstract:Adversarial Training (AT) improves neural network robustness, but most methods train a fixed parameter space from the start. This paper asks whether the order in which parameters become optimizable can affect the final robust solution, even when the final architecture or computation budget is controlled. We propose GRAPE, Guided Parameter-Space Evolution, a training framework for compact adversarial robustness. GRAPE combines parameter-space stabilization with progressive hidden expansion: it stabilizes robust optimization in the currently exposed space, gradually releases new optimizable dimensions, and uses an adversarial spectral utilization score to guide newly released capacity toward high-pressure modules. In contrast to fixed-structure AT, GRAPE treats robust model learning as a process of progressive parameter-space exposure and evolution. Under the standard $\ell_\infty$ threat model on CIFAR-10, with fixed-structure ResNet-18 AT as a controlled reference, GRAPE improves PGD-20 robust accuracy from 51.70% to 56.94% at a nearly matched computation budget with a FLOPs ratio of 1.009x, while reducing parameter count by about 21.4%. A sequential grow variant with the same final ResNet-18 architecture reaches 56.52% PGD-20 robust accuracy, indicating that the gain is not only due to final architecture differences but also to the parameter-space exposure path. These results suggest that guided parameter-space evolution can yield compact and robust parameter configurations under matched computation.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2606.14865 [cs.LG]
(or arXiv:2606.14865v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.14865 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 261 cuvinte
Subiecte conexe
Lattice-Based Cryptography Explained
Quantum Zeitgeist · 15 aug. 2026
The Alignment Problem: Computer Science Meets Analytic Philosophy
LessWrong · 18 oct. 2026
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space · 13 oct. 2026
Problems of Proliferation
LessWrong · 29 sept. 2026