AI Paper Atlas

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

Enhanced Accuracy and Robustness via Multi-Teacher Adversarial Distillation

Shiji Zhao, Jie Yu, Zhenlong Sun, Bo Zhang, Xingxing Wei

ECCV 2022 · Adversarial robustness and security › Certified and empirical robustness · approach: Adversarial training recipes · 49 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 Adaptive Normalization Loss to dynamically weight clean and adversarial teachers (Baseline+MT+ANL) (ablation)Baseline+MT · CIFAR-10 ablation (Figure 2)W-Robust accuracyImproved
add a clean teacher alongside the adversarial teacher with fixed 0.5 weights (Baseline+MT) (ablation)Baseline · CIFAR-10 ablation (Figure 2)clean and robust accuracyMixed
add Adaptive Normalization Loss: training loss less oscillating and relative loss gap smaller (ablation)Baseline+MT · CIFAR-10 training loss (Figures 3, 4)training lossImproved
add clean teacher to adversarial distillation (MTARD vs RSLAD), clean accuracyRSLAD · CIFAR-10Clean accuracy: 83.99% -> 87.36%Improved
add clean teacher to adversarial distillation (MTARD vs RSLAD), clean accuracyRSLAD · CIFAR-100Clean accuracy: 58.25% -> 64.3%Improved
add clean teacher with adaptive weights (MTARD vs RSLAD), PGDsat robust accuracyRSLAD · CIFAR-10 white-box PGDsatRobust accuracy: 53.94% -> 50.73%Hurt
MTARD vs RSLAD under FGSM W-Robust accuracyRSLAD · CIFAR-10 white-box FGSMW-Robust accuracy: 72.2% -> 74.28%Improved
MTARD vs RSLAD under PGDsat W-Robust accuracyRSLAD · CIFAR-100 white-box PGDsatW-Robust accuracy: 44.72% -> 44.63%No clear effect
MTARD vs RSLAD under CW-infinity W-Robust accuracyRSLAD · CIFAR-10 white-box CWW-Robust accuracy: 68.33% -> 67.97%Hurt
MTARD vs ARD under Square attack W-Robust accuracyARD · CIFAR-10 black-box Square attackW-Robust accuracy: 78.71% -> 83.68%Improved
MTARD vs ARD under PGD-20 transfer attack W-Robust accuracyARD · CIFAR-100 black-box PGD-20 transferW-Robust accuracy: 49.82% -> 52.85%Improved
use adversarial distillation (ARD, RSLAD, MTARD) instead of SAT and TRADESsmall student models · CIFAR-10/100 white-boxW-Robust 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