AI Paper Atlas

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

Concise Explanations of Neural Networks using Adversarial Training

Prasad Chalasani, Jiefeng Chen, Amrita Roy Chowdhury, Xi Wu, Somesh Jha

ICML 2020 · Adversarial robustness and security › Certified and empirical robustness · approach: Decision-boundary geometry of robustness · 30 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 ℓ∞-bounded adversarial training instead of natural training with ℓ1 regularizationCNNs · MNISTIG attribution sparseness improvement (dG): 0.004 vs 0.06Improved
Use ℓ∞-bounded adversarial training instead of natural training with ℓ1 regularizationCNNs · Fashion-MNISTattribution sparseness improvement (dG): 0.008 vs 0.06; 0.003 vs 0.08Improved
Use ℓ∞-bounded adversarial training instead of natural training with ℓ1 regularizationCNN · CIFAR-10IG attribution sparseness improvement (dG): 0.022 vs 0.081Improved
Use ℓ∞-bounded adversarial training instead of natural training with ℓ1 regularizationlogistic regression · MushroomIG attribution sparseness improvement (dG): 0.06 vs 0.06No clear effect
Use ℓ∞-bounded adversarial training instead of natural training with ℓ1 regularizationlogistic regression · SpambaseIG attribution sparseness improvement (dG): 0.15 vs 0.17Improved
Use ℓ∞-bounded adversarial training1-layer models with the specified loss and data assumptions · theoretical analysisIG attribution sparsenessImproved
Add an IG-stability penalty to natural training1-layer models with the specified loss · theoretical analysisadversarial objective equivalenceNo 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