AI Paper Atlas

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

Reliable Adversarial Distillation with Unreliable Teachers

Jianing Zhu, Jiangchao Yao, Bo Han, Jingfeng Zhang, Tongliang Liu, Gang Niu, Jingren Zhou, Jianliang Xu, Hongxia Yang

ICLR 2022 · Adversarial robustness and security › Attacks on text and graphs · approach: Defenses against universal perturbations · 20 citations (OpenAlex)

Read the paper

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
Use IAD-I instead of ARDadversarial distillation with ResNet-18 and TRADES teacher · CIFAR-10PGD-20 accuracy: 53.38% -> 53.84%Improved
Use IAD-I instead of ARDadversarial distillation with ResNet-18 and TRADES teacher · CIFAR-100PGD-20 accuracy: 30.21% -> 30.60%Improved
Add student self-introspection to AKD² (ablation)IAD-II, ResNet-18 · CIFAR-10PGD-20 accuracy: 51.36% -> 51.85%Improved
Increase β from 0.01 to 1.0, raising student introspection (ablation)IAD-I, ResNet-18 · CIFAR-10PGD-20 accuracy: 51.49% -> 52.64%Improved
Increase β from 0.01 to 1.0, raising student introspection (ablation)IAD-I, ResNet-18 · CIFAR-10Natural accuracy: 82.76% -> 80.79%Hurt
Use 60-epoch rather than no warm-up (ablation)IAD-I, ResNet-18 · CIFAR-10PGD-20 accuracy: 51.75% -> 52.14%Improved
Use 60-epoch rather than no warm-up (ablation)IAD-II, ResNet-18 · CIFAR-10PGD-20 accuracy: 52.07% -> 52.08%No clear effect
Downweight teacher KL guidance by α instead of using constant weight 1.0 (ablation)IAD-I-Down versus IAD-I · CIFAR-10PGD-20 accuracy: 52.14% -> 51.77%Hurt
Downweight teacher KL guidance by α instead of using constant weight 1.0 (ablation)IAD-I-Down versus IAD-I · CIFAR-100PGD-20 accuracy: 27.40% -> 27.32%Hurt
Downweight teacher KL guidance by α instead of using constant weight 1.0 (ablation)IAD-I-Down versus IAD-I · Tiny-ImageNetPGD-20 accuracy: 23.40% -> 22.68%Hurt
Increase the student-introspection coefficient from 1 - α to 4.0 (ablation)IAD-I, ResNet-18 · CIFAR-10PGD-20 accuracy: 52.14% -> 55.78%Improved
Increase the student-introspection coefficient from 1 - α to 4.0 (ablation)IAD-I, ResNet-18 · CIFAR-10Natural accuracy: 82.09% -> 77.34%Hurt

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