Research map / Adversarial robustness and security / Certified and empirical robustness
Excessive Invariance Causes Adversarial Vulnerability
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Add independence cross-entropy to training instead of using vanilla cross-entropy (ablation) | Fully invertible RevNet · MNIST test | logit classifier error (%): 0.39 -> 0.38 | No clear effect |
| Add independence cross-entropy to training instead of using vanilla cross-entropy (ablation) | Fully invertible RevNet · MNIST train | logit classifier error (%): 0.00 -> 0.37 | Mixed |
| Add independence cross-entropy to training instead of using vanilla cross-entropy (ablation) | Fully invertible RevNet · MNIST test | nuisance-classifier error (%): 0.34 -> 27.70 | Helped |
| Add independence cross-entropy to training instead of vanilla cross-entropy (ablation) | Fully invertible RevNet · Binary shiftMNIST | error reduction: up to 38% | Helped |
| Add independence cross-entropy to training instead of vanilla cross-entropy (ablation) | Fully invertible RevNet · Textured shiftMNIST | error reduction: around 28% | Helped |
| Train with vanilla cross-entropy rather than independence cross-entropy (ablation) | Fully invertible RevNet · MNIST metameric sampling | preservation of class semantics under nuisance interpolation | Hurt |
| Add independence cross-entropy rather than use vanilla cross-entropy (ablation) | Fully invertible RevNet · MNIST metameric sampling | preservation of class semantics under nuisance interpolation | Helped |
Closest papers by meaning
- Rethinking Invariance Regularization in Adversarial Training to Improve Robustness-Accuracy Trade-off (ICLR 2025 · similarity 0.88)
- Contrasting Adversarial Perturbations: The Space of Harmless Perturbations (AAAI 2025 · similarity 0.87)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.86)
- Doppelgangers and Adversarial Vulnerability (CVPR 2025 · similarity 0.86)
- Where and How to Attack? A Causality-Inspired Recipe for Generating Counterfactual Adversarial Examples (AAAI 2024 · similarity 0.86)
- Layer-Aware Analysis of Catastrophic Overfitting: Revealing the Pseudo-Robust Shortcut Dependency (ICML 2024 · similarity 0.86)
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.