AI Paper Atlas

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

Adversarial Margin Maximization Networks

Ziang Yan, Yiwen Guo, Changshui Zhang

TPAMI 2019 · Adversarial robustness and security › Certified and empirical robustness · approach: Augmentation and diffusion purification · 14 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
use MIN aggregation instead of AVG with LIN shrinkage (ablation)AMM on MLP · MNISTerror rate (%): 2.26 -> 1.28Improved
add AVG+LIN regularizer to the MLP (ablation)MLP baseline · MNISTerror rate (%): 1.79 -> 2.26Hurt
use EXP shrinkage instead of LIN under AVG (ablation)AMM on MLP · MNISTerror rate (%): 2.26 -> 0.94Improved
use INV shrinkage instead of LIN under AVG (ablation)AMM on MLP · MNISTerror rate (%): 2.26 -> 1.18Improved
use EXP shrinkage instead of LIN under MIN (ablation)AMM on MLP · MNISTerror rate (%): 1.28 -> 0.90Improved
use MIN instead of AVG aggregation with EXP (ablation)AMM on MLP · MNISTerror rate (%): 0.94 -> 0.90Improved
back-propagate through grad f (high order gradients) (ablation)AMM on MLP · MNISTerror rate (%): 1.46 -> 0.90Improved
add AMM to MLP trained on 20k images vs vanilla on 60kMLP · MNISTerror rate (%): 1.81 -> 1.54Improved
add AMM when 55k of 60k labels are randomMLP · MNIST noisy labelserror rate (%): above 40% -> below 10%Improved
add AMM to LeNetLeNet · MNISTerror rate (%): 0.87 -> 0.56Improved
add AMM to NINNIN · CIFAR-100error rate (%): 37.75 -> 34.49Improved
add AMM to ResNet-20ResNet-20 · CIFAR-100error rate (%): 33.20 -> 32.96No 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