Research map / Optimization / Implicit bias and generalization
Toward Better Generalization Bounds with Locally Elastic Stability
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 expected data-dependent sensitivity instead of worst-case sensitivity | Neural networks · CIFAR-10 | sensitivity: 3.05 -> 0.02 | Helped |
| Use expected data-dependent sensitivity instead of worst-case sensitivity | Random feature model · CIFAR-10 | sensitivity: 1.73 -> 0.04 | Helped |
| Use locally elastic stability bounds instead of uniform stability bounds | Bounded SVM regression · Bounded SVM regression under the stated input-distribution conditions | generalization bound | Helped |
| Use locally elastic stability bounds instead of uniform stability bounds | Regularized least square regression · Regularized least square regression under the stated input-distribution conditions | generalization bound | Helped |
| Use locally elastic stability instead of uniform stability | SGD · Non-convex optimization under the stated conditions | generalization bound | Helped |
Closest papers by meaning
- Algorithmic Stability Unleashed: Generalization Bounds with Unbounded Losses (ICML 2024 · similarity 0.91)
- Fantastic Generalization Measures are Nowhere to be Found (ICLR 2024 · similarity 0.90)
- Uniform Stability and Generalization Error of GD and SGD on Fixed-Point Parameters (NeurIPS 2026 · similarity 0.89)
- Certification-Enhanced Generalization Bounds (NeurIPS 2026 · similarity 0.89)
- Generalization Analysis of Biased Stochastic Gradient Methods for Minimax Problems (NeurIPS 2026 · similarity 0.88)
- Stability and Generalization Analysis of Decentralized SGD: Sharper Bounds Beyond Lipschitzness and Smoothness (ICML 2025 · similarity 0.88)
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.