AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Do Input Gradients Highlight Discriminative Features?

Harshay Shah, Prateek Jain, Praneeth Netrapalli

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Self-explaining and visual explanation networks · 1 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
train with l2/l_inf adversarial robustness instead of standard training (ablation)MLPs and Resnets · SVHN, Fashion MNIST, CIFAR-10, ImageNet-10DiffROAR: -5 -> 15Improved
adversarially robust MLPs vs standard MLPs at k=10% (ablation)MLP · SVHNDiffROAR: roughly 10-30%Improved
standard Resnets on CIFAR-10 and ImageNet-10 at k<40%standard Resnet · CIFAR-10, ImageNet-10DiffROARHurt
adversarial training of SVHN MLP at k=5% (top-k vs bottom-k predictive power) (ablation)MLP, l_inf eps=2/255 · SVHNpredictive power: roughly 40%Improved
adversarial robust model vs standard model on BlockMNIST attributions (ablation)MLP and Resnet18 · BlockMNISTattribution on signal vs null blockImproved
lower learning rate 0.1 to 0.01, weight decay 0.0001 to 0, momentum 0.9 to 0 (ablation)retraining on masked data · CIFAR-10DiffROAR: 0.1 -> 0.01No clear effect
increase PGD adversarial training perturbation budget epsilon (ablation)adversarially trained models · CIFAR-10DiffROARImproved
evaluate BlockMNIST-Top-trained models on BlockMNIST-BottomResnet · BlockMNISTtest accuracy (%): 100 -> 55Hurt
adversarial training increases robustness but lowers clean accuracyMLPs and CNNs · SVHN, Fashion MNIST, CIFAR-10, ImageNet-10standard test accuracyHurt

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