Research map / Learning theory and algorithms / Kernels and regression theory
How many samples are needed to leverage smoothness?
Vivien Cabannes, Stefano Vigogna
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 samples n past the effective dimension N_1(sigma, lambda) in kernel ridge regression | regularized kernel ridge regression · f*(x)=sign(<x,e1>), uniform on [-1,1]^2 | excess risk | Mixed |
| tune hyperparameters so that effective dimension is smaller than n | kernel ridge regression · synthetic regression | excess risk peak | Improved |
| decrease regularization lambda from 10^6 to 10^2 to 1 on the torus | Gaussian kernel with sigma=1 · torus R^2/Z^2 weights alpha_x(x0) | spatial extent of weights | Mixed |
| vary n in {27, 45, 90, 180, 360, 720} relative to N_2(K)=45 | Gaussian kernel ridge regression, sigma=.1, lambda=10^-5 · X=[-1,1], uniform | learned weights behavior | Hurt |
Closest papers by meaning
- Functions with average smoothness: structure, algorithms, and learning (JMLR 2024 · similarity 0.86)
- Near-optimal learning with average Hölder smoothness (NeurIPS 2023 · similarity 0.85)
- Learnability of convolutional neural networks for infinite dimensional input via mixed and anisotropic smoothness (ICLR 2022 · similarity 0.85)
- Algorithmic Stability Unleashed: Generalization Bounds with Unbounded Losses (ICML 2024 · similarity 0.82)
- Scale-Sensitive Shattering: Learnability and Evaluability at Optimal Scale (NeurIPS 2026 · similarity 0.82)
- Certification-Enhanced Generalization Bounds (NeurIPS 2026 · 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.
