Research map / Adversarial robustness and security / Certified and empirical robustness
Rademacher Complexity for Adversarially Robust Generalization
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 |
|---|---|---|---|
| Increase ℓ1 regularization coefficient λ | MNIST multi-class linear classifier trained adversarially · MNIST | adversarial generalization gap | Helped |
| Increase feature dimension | MNIST multi-class linear classifier trained adversarially · MNIST image features | adversarial generalization gap | Hurt |
| Increase ℓ1 regularization coefficient λ | Four-layer ReLU neural network trained with PGD adversarial training · MNIST | adversarial generalization error | Helped |
| Use an ℓ1-bounded weight vector instead of only an ℓp bound with p > 1 | Binary linear classifier under an ℓ∞ attack · Adversarial Rademacher complexity | dimension dependence | Helped |
| Use an ℓ1 constraint on the first-layer weight matrix | One-hidden-layer ReLU network with surrogate adversarial loss · Adversarial generalization bound | dimension dependence | Helped |
Closest papers by meaning
- Adversarial Learning Guarantees for Linear Hypotheses and Neural Networks (ICML 2020 · similarity 0.96)
- Adversarially Robust Generalization Requires More Data (NeurIPS 2018 · similarity 0.90)
- PAC-Bayesian Spectrally-Normalized Bounds for Adversarially Robust Generalization (NeurIPS 2023 · similarity 0.89)
- On the Generalization Analysis of Adversarial Learning (ICML 2022 · similarity 0.89)
- Why Robust Generalization in Deep Learning is Difficult: Perspective of Expressive Power (NeurIPS 2022 · similarity 0.89)
- $\sigma$-zero: Gradient-based Optimization of $\ell_0$-norm Adversarial Examples (ICLR 2025 · similarity 0.87)
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.