Research map / Learning theory and algorithms / Kernels and regression theory
Gaussian Quadrature for Kernel Features
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 |
|---|---|---|---|
| Use subsampled dense-grid features instead of random Fourier features | Sparse ANOVA kernel classification · MNIST | test accuracy | Helped |
| Use subsampled dense-grid features instead of random Fourier features | Sparse ANOVA kernel classification · MNIST | feature generation time: 320 seconds vs. 384 seconds; a savings of 17% | Helped |
| Use subsampled dense-grid features instead of random Fourier features | Sparse ANOVA kernel acoustic modeling · TIMIT | phone error rate | No clear effect |
| Use reweighted quadrature features instead of random Fourier features | Kernel approximation · TIMIT dev set | root mean squared error: 2--3 times lower error | Helped |
| Use reweighted quadrature features instead of random Fourier features | Kernel approximation · TIMIT dev set | number of features at a fixed approximation-error threshold: 3--5 times fewer features | Helped |
| Use reweighted quadrature features instead of random Fourier features | Sparse ANOVA kernel acoustic modeling · TIMIT | phone error rate | Helped |
| Use quadrature-based features instead of random Fourier features | Kernel feature generation · TIMIT | feature generation time: about twice as fast | Helped |
| Use polynomially-exact quadrature instead of random Fourier features | Gaussian kernel approximation · 25-dimensional Gaussian kernel; small-diameter regions | maximum kernel approximation error | Helped |
| Use sparse-grid quadrature instead of random Fourier features | Gaussian kernel approximation · 25-dimensional Gaussian kernel; small-diameter regions | maximum kernel approximation error | Helped |
| Use sparse-grid quadrature instead of random Fourier features | Gaussian kernel approximation · 25-dimensional Gaussian kernel; larger region diameters | maximum kernel approximation error | Hurt |
| Use subsampled dense-grid features instead of random Fourier features | Kernel approximation · TIMIT dev set | root mean squared error | No clear effect |
Closest papers by meaning
- Quadrature-based features for kernel approximation (NeurIPS 2018 · similarity 0.90)
- Towards a Unified Analysis of Random Fourier Features (ICML 2019 · similarity 0.89)
- Large-Scale Learning with Fourier Features and Tensor Decompositions (NeurIPS 2021 · similarity 0.88)
- Random Features for Shift-Invariant Kernels with Moment Matching (AAAI 2017 · similarity 0.88)
- A random matrix analysis of random Fourier features: beyond the Gaussian kernel, a precise phase transition, and the corresponding double descent (NeurIPS 2020 · similarity 0.87)
- Random features for Grassmannian kernel approximation with bounded rank-one projections (NeurIPS 2026 · similarity 0.86)
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.