Research map / Adversarial robustness and security / Certified and empirical robustness
Lipschitz Bounds and Provably Robust Training by Laplacian Smoothing
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 the Lipschitz bound alpha (ablation) | Lipschitz-constrained graph classifier · MNIST (n=5000) | test accuracy and confidence | Helped |
| increase the Lipschitz bound alpha (ablation) | Lipschitz-constrained graph classifier · MNIST (n=5000) | sensitivity to perturbations | Hurt |
| increase number of graph vertices at alpha=300 (ablation) | Lipschitz-constrained graph classifier · MNIST | test accuracy and loss | Helped |
| increase number of vertices (ablation) | Lipschitz-constrained graph classifier · MNIST (alpha=300) | test confidence | No clear effect |
| increase loss margin epsilon from below 0.25 to 0.25 (ablation) | Loss-constrained Lipschitz minimization · Toy 2D classification | test accuracy: 0.93 -> 0.54 | Hurt |
| increase loss margin epsilon below 0.25 (ablation) | Loss-constrained Lipschitz minimization · Toy 2D classification | test accuracy: 0.93 | No clear effect |
| increase loss margin epsilon up to 0.25 (ablation) | Loss-constrained Lipschitz minimization · Toy 2D classification | test confidence: decreases to 0.5 | Hurt |
| increase loss margin epsilon (ablation) | Loss-constrained Lipschitz minimization · MNIST (n=5000) | Lipschitz constant, accuracy, confidence | Hurt |
| increase graph nodes n (125, 200, 500) and Lipschitz bound (ablation) | Lipschitz-constrained graph classifier · Toy 2D checkerboard | classification confidence | Helped |
Closest papers by meaning
- Improved techniques for deterministic l2 robustness (NeurIPS 2022 · similarity 0.91)
- A Recipe for Improved Certifiable Robustness (ICLR 2024 · similarity 0.89)
- ECLipsE: Efficient Compositional Lipschitz Constant Estimation for Deep Neural Networks (NeurIPS 2024 · similarity 0.88)
- Efficient Robust Conformal Prediction via Lipschitz-Bounded Networks (ICML 2025 · similarity 0.88)
- Some Fundamental Aspects about Lipschitz Continuity of Neural Networks (ICLR 2024 · similarity 0.87)
- Delving into the Convergence of Generalized Smooth Minimax Optimization (ICML 2024 · 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.