Research map / Learning theory and algorithms / Kernels and regression theory
Fast Exact Evaluation of Univariate Kernel Sums
David P. Hofmeyr
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 |
|---|---|---|---|
| use exact K1 kernel instead of Gaussian approximation | Truncated Taylor Gaussian · Density (c) grid, n=1e3 | MISE: 1.02e-01 -> 8.33e-02 | Improved |
| use exact K1 kernel instead of Gaussian approximation | Truncated Taylor Gaussian · Density (c) grid, n=1e6 | MISE: 3.73e-03 -> 3.54e-03 | Improved |
| use K4 instead of Gaussian kernel | Exact Gaussian · Density (c) grid, n=1e3 | MISE: 1.02e-01 -> 1.05e-01 | No clear effect |
| use K1 kernel at sample points | Truncated Taylor Gaussian · Density (c) sample points, n=1e5 | ISE: 1.58e-02 -> 1.37e-02 | Improved |
| use exact K1 instead of truncated Taylor approximation | Truncated Taylor Gaussian · Density (d) sample points, n=1e5 | running time (s): 3.45e-01 -> 5.82e-02 | Improved |
| use exact K4 instead of truncated Taylor approximation | Truncated Taylor Gaussian · Density (d) sample points, n=1e5 | running time (s): 3.45e-01 -> 2.04e-01 | Improved |
| use recursive kernel sums instead of exact Gaussian | Exact Gaussian · Density (d) sample points, n=1e3 | running time (s): 1.86e-01 -> 7e-04 | Improved |
| use K1 for first derivative estimation | Truncated Taylor Gaussian · Density (c) derivative, n=1e5 | ISE: 8.45e+00 -> 7.35e+00 | Improved |
| use lower-efficiency K1 on smooth density (a) derivative | Truncated Taylor Gaussian · Density (a) derivative, n=1e5 | ISE: 3.71e-04 -> 5.12e-04 | Hurt |
| use K7 instead of K1 for first derivative | Exact K1 · Density (c) derivative, n=1e5 | ISE: 7.35e+00 -> 9.55e+00 | Hurt |
| use exact K1 for first derivative | Truncated Taylor Gaussian · Density (a) derivative, n=1e5 | running time (s): 5.74e-01 -> 5.82e-02 | Improved |
| use binned K1 instead of exact K1 | Exact K1 · Density (c) grid, n=1e6 | MISE: 3.54e-03 -> 3.57e-03 | No clear effect |
Closest papers by meaning
- Even Faster Kernel Matrix Linear Algebra via Density Estimation (ICML 2026 · similarity 0.88)
- SD-KDE: Score-Debiased Kernel Density Estimation (NeurIPS 2025 · similarity 0.87)
- Faster Kernel Matrix Algebra via Density Estimation (ICML 2021 · similarity 0.87)
- Uniform Convergence Rates for Kernel Density Estimation (ICML 2017 · similarity 0.87)
- Adaptive Greedy Algorithm for Moderately Large Dimensions in Kernel Conditional Density Estimation (JMLR 2022 · similarity 0.87)
- Dynamic Similarity Graph Construction with Kernel Density Estimation (ICML 2025 · similarity 0.84)
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.
