Research map / Learning theory and algorithms / Combinatorial and search optimization
Learning Optimal Feedback Operators and their Sparse Polynomial Approximations
Karl Kunisch, Donato Vásquez-Varas, Daniel Walter
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 |
|---|---|---|---|
| increase number of training initial conditions (ablation) | learned feedback law · Cucker-Smale optimal consensus | test error: cardinalities 5, 10, 15 and 20; errors smaller than 2.1 % | Improved |
| train with two or more initial conditions instead of one (ablation) | learned feedback law · Modified Van der Pol oscillator | stabilization of test initial conditions: one initial condition fails near right-hand branch; two or more steer all to origin | Improved |
| increase training set size from 1 to 2 (ablation) | learned feedback law · LC-circuit | objective error bars: Ntrain = 1 error-bars distinctly larger than Ntrain = 2 and 10 | Improved |
| increase polynomial degree (ablation) | learned feedback law · Modified Van der Pol oscillator | support cardinality and solve time: degrees 4 to 8; closed-loop solve time under 0.42 seconds | Hurt |
| increase sparsity penalty gamma from 1e-1 to 1e-3 (ablation) | learned feedback law · Allen-Cahn equation | MNSE_J: slope 0.6 and intercept 0.13 at gamma 1e-1 to closer to identity at 1e-3 | Improved |
| increase gamma (ablation) | learned feedback law · Allen-Cahn equation | support size and training time: support size decreases with gamma over 10^-1 to 10^-6 | Improved |
| increase training cardinality from 5 to 20 (ablation) | learned feedback law · Allen-Cahn (gamma 1e-3) | regression to identity line: cardinality 5 to 20 | Improved |
| use full gradient (38) instead of greedy (40) in Algorithm 1 (ablation) | learned feedback law · Cucker-Smale optimal consensus | error and training time: (40) gave larger errors and training times but sparser solutions | Mixed |
Closest papers by meaning
- Constrained Optimization From a Control Perspective via Feedback Linearization (NeurIPS 2025 · similarity 0.86)
- Learning-enabled Polynomial Lyapunov Function Synthesis via High-Accuracy Counterexample-Guided Framework (CVPR 2025 · similarity 0.83)
- Optimal Control Operator Perspective and a Neural Adaptive Spectral Method (AAAI 2025 · similarity 0.83)
- Learning to Relax: Setting Solver Parameters Across a Sequence of Linear System Instances (ICLR 2024 · similarity 0.82)
- Physics-Informed Neural Network Policy Iteration: Algorithms, Convergence, and Verification (ICML 2024 · similarity 0.82)
- Mean-Field Analysis for Learning Subspace-Sparse Polynomials with Gaussian Input (NeurIPS 2024 · similarity 0.82)
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.
