Research map / Learning theory and algorithms / Combinatorial and search optimization
Semialgebraic Optimization for Lipschitz Constants of ReLU Networks
Tong Chen, Jean B Lasserre, Victor Magron, Edouard Pauwels
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 changed | Where | Measured | Effect |
|---|---|---|---|
| HR-2 SDP relaxation instead of LipOpt-3 LP | (80,80) network · global Lipschitz bound, sparsity 80 | Lipschitz upper bound (lower better): 4.68 -> 2.68 | Improved |
| HR-2 instead of LipOpt-3 at low sparsity | (80,80) network · global Lipschitz bound, sparsity 20 | Lipschitz upper bound: 1.55 -> 1.45 | Mixed |
| HR-2 instead of LipOpt-3 | (80,80) network · solver time, sparsity 80 | time (s): 71.49 -> 9.82 | Improved |
| HR-2 instead of Shor relaxation | SDP-NN (784,500) · global Lipschitz bound | Lipschitz upper bound: 17.85 -> 14.56 | Improved |
| HR-2 instead of Shor relaxation | SDP-NN (784,500) · solver time | time (s): 2869 -> 12246 | Hurt |
| HR-2 instead of LipOpt-3 for certification | (80,80) binary classifier · certified ratio eps=0.03 | certified ratio: 4.78% -> 62.46% | Improved |
| HR-2 instead of LipOpt-4 | (80,80) network · global Lipschitz bound, sparsity 20 | Lipschitz upper bound: 1.43 -> 1.45 | Mixed |
Closest papers by meaning
- ECLipsE: Efficient Compositional Lipschitz Constant Estimation for Deep Neural Networks (NeurIPS 2024 · similarity 0.89)
- Novel Quadratic Constraints for Extending LipSDP beyond Slope-Restricted Activations (ICLR 2024 · similarity 0.89)
- Characterizing ResNet's Universal Approximation Capability (ICML 2024 · similarity 0.88)
- Convex Relaxations of ReLU Neural Networks Approximate Global Optima in Polynomial Time (ICML 2024 · similarity 0.88)
- Approximation theory for 1-Lipschitz ResNets (NeurIPS 2025 · similarity 0.87)
- Some Fundamental Aspects about Lipschitz Continuity of Neural Networks (ICLR 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.
