Research map / Learning theory and algorithms / Kernels and regression theory
PAC-Bayes bounds for stable algorithms with instance-dependent priors
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 the instance-dependent PAC-Bayes bound centered at the expected SVM weight (PEW) instead of the prior-at-origin bound (PO) | Gaussian-randomized SVM · Pima Indians (PIM), small C (C ≤ 0.5) | bound tightness | Helped |
| Use the instance-dependent PEW bound instead of the PO bound | Gaussian-randomized SVM · RIN | bound tightness | Mixed |
| Select SVM parameters using the PEW bound rather than the PO bound | Gaussian-randomized SVM · RIN | test error: around 0.12 | Helped |
| Use either of the compared hinge-loss-based stability bounds | SVM · PIM and RIN | bound tightness | Hurt |
| Increase C when using stability-based bounds | PEW and other stability-based SVM bounds · PIM and RIN | bound tightness | Hurt |
Closest papers by meaning
- PAC-Bayes Learning Bounds for Sample-Dependent Priors (NeurIPS 2020 · similarity 0.91)
- Toward Better PAC-Bayes Bounds for Uniformly Stable Algorithms (NeurIPS 2023 · similarity 0.91)
- PAC-Bayes Analysis Beyond the Usual Bounds (NeurIPS 2020 · similarity 0.90)
- Controlling Multiple Errors Simultaneously with a PAC-Bayes Bound (NeurIPS 2024 · similarity 0.88)
- PAC-Bayes-Chernoff bounds for unbounded losses (NeurIPS 2024 · similarity 0.87)
- Recursive PAC-Bayes: A Frequentist Approach to Sequential Prior Updates with No Information Loss (NeurIPS 2024 · similarity 0.87)
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.