Research map / Learning theory and algorithms / Kernels and regression theory
Time-Accuracy Tradeoffs in Kernel Prediction: Controlling Prediction Quality
Samory Kpotufe, Nakul Verma
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 |
|---|---|---|---|
| raise knob alpha from 1/6 to 6/6 | r-nets kernel prediction · CT Slices (51k) | error ratio: 0.93 -> 0.43 | Hurt |
| raise knob alpha from 1/6 to 6/6 | r-nets kernel prediction · CT Slices (51k) | time ratio (speedup): 1.29 -> 15.33 | Improved |
| raise knob alpha from 1/6 to 6/6 | r-nets kernel prediction · SARCOS (42k) | error ratio: 0.99 -> 0.77 | Hurt |
| raise knob alpha from 1/6 to 6/6 | r-nets kernel prediction · SARCOS (42k) | time ratio (speedup): 2.03 -> 10.14 | Improved |
| raise knob alpha from 1/6 to 6/6 | r-nets kernel prediction · MiniBooNE (128k) | error ratio: 0.99 -> 0.88 | Hurt |
| raise knob alpha from 1/6 to 6/6 | r-nets kernel prediction · MiniBooNE (128k) | time ratio (speedup): 1.17 -> 10.22 | Improved |
| raise alpha from 1/6 to 4/6 | r-nets kernel prediction · SARCOS (42k) | error ratio / time ratio: 0.99 - 2.03 -> 0.96 - 7.70 | Mixed |
| raise alpha from 1/6 to 4/6 | r-nets kernel prediction · MiniBooNE (128k) | error ratio / time ratio: 0.99 - 1.17 -> 0.98 - 6.42 | Mixed |
| replace r-net quantization with kd-tree quantization (level 15) (ablation) | r-nets kernel prediction · CT Slices (51k) | error ratio / time ratio: 0.91 - 5.40 -> 0.41 - 2.89 | Hurt |
| replace r-net quantization with kd-tree quantization (level 15) (ablation) | r-nets kernel prediction · SARCOS (42k) | error ratio / time ratio: 0.93 - 1.42 | Hurt |
| increase sample size n | r-nets kernel prediction · r-nets, all datasets | accuracy and speedup | Improved |
Closest papers by meaning
- On the Target-kernel Alignment: a Unified Analysis with Kernel Complexity (NeurIPS 2024 · similarity 0.88)
- Generalization in Kernel Regression Under Realistic Assumptions (ICML 2024 · similarity 0.86)
- Data-centric Prediction Explanation via Kernelized Stein Discrepancy (ICLR 2025 · similarity 0.85)
- Supervised Kernel Thinning (NeurIPS 2024 · similarity 0.85)
- The optimality of kernel classifiers in Sobolev space (ICLR 2024 · similarity 0.85)
- Kernel Representation Learning for Time Sequence: Algorithm, Theory, and Applications (AAAI 2025 · similarity 0.85)
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.
