AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

GradAuto: Energy-Oriented Attack on Dynamic Neural Networks

Jianhong Pan, Qichen Zheng, Zhipeng Fan, Hossein Rahmani, Qiuhong Ke, Jun Liu

ECCV 2022 · Adversarial robustness and security › Attacks on text and graphs · approach: Attacks and defenses in sequential settings · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Rectified Direction (gradient projection removing conflict between Complexity and Finished gradients) (ablation)ILFO · ImageNet SkipNet+HRLComplexity increase (%): 46.9 -> 52.3Improved
add Rectified Magnitude (Lipschitz-bounded gradient magnitude) (ablation)ILFO · ImageNet SkipNet+HRLComplexity increase (%): 46.9 -> 86.9Improved
add Rectified Direction on top of Rectified Magnitude (ablation)Rectified Magnitude · ImageNet SkipNet+HRLComplexity increase (%): 86.9 -> 88.9Improved
replace ILFO with GradAuto on SkipNet+HRLILFO · ImageNet SkipNet+HRLFLOPs recovery (%): 46.9 -> 88.9Improved
replace ILFO with GradAuto on SkipNet, CIFAR-10ILFO · CIFAR-10 SkipNetFLOPs recovery (%): 84.29 -> 100Improved
replace ILFO with GradAuto on SACT, CIFAR-10ILFO · CIFAR-10 SACTFLOPs recovery (%): 72.49 -> 100Improved
replace ILFO with GradAuto on ManiDP, ImageNetILFO · ImageNet ManiDPFLOPs recovery (%): 85.6 -> 100Improved
lower the Lipschitz constant from 1e+0 to 1e-3 (ablation)GradAuto · ImageNet SkipNet+HRLFLOPs recovery (%): 88.9 -> 71.0Mixed
raise Lipschitz constant from 1e-3 to 1e+0 (ablation)GradAuto · ImageNet SkipNet+HRLFLOPs recovery (%): 71.0 -> 88.9Mixed
raise Lipschitz constant beyond 1e+0 (to 1e+2) (ablation)GradAuto · ImageNet SkipNet+HRLFLOPs recovery (%): 88.9 -> 88.9No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy