AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Learning Extremely Lightweight and Robust Model with Differentiable Constraints on Sparsity and Condition Number

Xian Wei, Yangyu Xu, Yanhui Huang, Hairong Lv, Hai Lan, Mingsong Chen, Xuan Tang

ECCV 2022 · Adversarial robustness and security › Certified and empirical robustness · approach: Robustness-accuracy trade-off analysis · 3 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 condition number and sparsity constraints with tensor product (ARLST) vs ATMCVGG-16 · CIFAR-10 PGD, 200xrobust accuracy: 36.1 -> 37.5Improved
ARLST instead of HYDRA at 200x compressionVGG-16 · CIFAR-10 PGD, 200xrobust accuracy: 35.5 -> 37.5Improved
ARLST instead of ADMM at 200x compressionVGG-16 · CIFAR-10 PGD, 200xrobust accuracy: 30.5 -> 37.5Improved
ARLST instead of ATMC at 100x compressionVGG-16 · CIFAR-10 PGD, 100xrobust accuracy: 41.1 -> 41.3Improved
ARLST instead of HYDRA at low compression (10x)VGG-16 · CIFAR-10 PGD, 10xrobust accuracy: 46.2 -> 46.3No clear effect
ARLST instead of ATMC on ImageNet-1K 10xVGG-16 · ImageNet-1K, 10xrobust accuracy: 27.8 -> 28.7Improved
ARLST instead of ATMC under PGD on CIFAR-100 (100x)VGG-16 · CIFAR-100 PGD, 100xrobust accuracy: 21.3 -> 22.1Improved
ARLST instead of ATMC under AutoAttack on CIFAR-100 (100x)VGG-16 · CIFAR-100 AutoAttack, 100xrobust accuracy: 19.9 -> 20.8Improved
ARLST instead of HYDRA on CCT at 10x compressionCCT · CIFAR-10 FGSM, 10xrobust accuracy: 36.45 -> 56.98Improved
ARLST instead of HYDRA on CCT at 5x compressionCCT · CIFAR-10 FGSM, 5xrobust accuracy: 55.54 -> 61.04Improved
add sparsity regularizer g (mu3 = 0.00001) to separable matrices (ablation)ARLST VGG-16 · CIFAR-10 (last FC layer weights)weight sparsityImproved
set condition number weights mu1 = 0.0001 and mu2 = 1 (ablation)ARLST · CIFAR-10 PGDrobust accuracyImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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