AI Paper Atlas

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

SAFER: Sharpness Aware layer-selective Finetuning for Enhanced Robustness in vision transformers

Bhavna Gopal, Huanrui Yang, Mark Horton, Yiran Chen

ICCV 2025 · Adversarial robustness and security › Certified and empirical robustness · approach: Adversarial training with feature consistency · 5 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
select layers by sharpness instead of randomly (ablation)SAFER on DeiT-Ti · CIFAR-10PGD-20 accuracy: 65.46 -> 68.50Improved
select layers by sharpness instead of randomly (ablation)SAFER on ViT-S · CIFAR-10Auto Attack accuracy: 46.19 -> 50.12Improved
select layers by sharpness instead of randomly (ablation)SAFER on Swin-B · CIFAR-10PGD-20 accuracy: 49.72 -> 53.65Improved
add SAFER layer-selective SAM finetuningPGD-AT (SGD) DeiT-Ti · CIFAR-10PGD-20 accuracy: 48.10 -> 68.50Improved
add SAFERPGD-AT (SGD) DeiT-Ti · CIFAR-10clean accuracy: 75.46 -> 82.36Improved
use SAFER instead of naive SAM in PGD-ATDeiT-S · CIFAR-10PGD-20 accuracy: 53.64 -> 70.26Improved
add SAFER on top of Tian et al.Tian et al. DeiT-Ti · CIFAR-10PGD-20 accuracy: 46.33 -> 48.05Improved
use CNN critical-layer method CLAT instead of PGD-ATDeiT-Ti · CIFAR-10PGD-20 accuracy: 48.10 -> 46.29Hurt
use RiFT instead of PGD-ATDeiT-Ti · CIFAR-10PGD-20 accuracy: 48.10 -> 44.32Hurt
use RiFT instead of PGD-ATViT-S · CIFAR-10PGD-20 accuracy: 50.86 -> 45.93Hurt
add SAFER with LoRALoRA PGD-AT (SAM) DeiT-Ti · CIFAR-10PGD-20 adversarial accuracy: 51.45 -> 63.50Improved
add SAFER on top of full PGD-AT after pretraining epochs (vs PGD-AT alone) (ablation)DeiT-Ti · CIFAR-10clean accuracyImproved

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