Research map / Optimization / Optimizers
A PAC-Bayesian Analysis of Randomized Learning with Application to Stochastic Gradient Descent
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 |
|---|---|---|---|
| adaptive sampling (multiplicative weight updates on sampling distribution) instead of uniform sampling | SGD with 0-1 utility (AdaSamp-01-SGD / AdaSamp-01-AdaGrad) · CIFAR-10 | empirical risk / training accuracy speed | Helped |
| adaptive sampling instead of uniform sampling | AdaGrad · CIFAR-10 | test accuracy in early iterations: modest, yet consistent | Helped |
| increase amplitude parameter alpha in adaptive sampling (ablation) | adaptive sampling SGD · CIFAR-10 | test accuracy | Hurt |
| adaptive sampling with multiplicative weight updates | SGD · CIFAR-10 | KL divergence from uniform prior | No clear effect |
Closest papers by meaning
- Toward Better PAC-Bayes Bounds for Uniformly Stable Algorithms (NeurIPS 2023 · similarity 0.92)
- $L_2$-Uniform Stability of Randomized Learning Algorithms: Sharper Generalization Bounds and Confidence Boosting (NeurIPS 2023 · similarity 0.89)
- Deterministic PAC-Bayesian generalization bounds for deep networks via generalizing noise-resilience (ICLR 2019 · similarity 0.89)
- On Generalization Error Bounds of Noisy Gradient Methods for Non-Convex Learning (ICLR 2020 · similarity 0.89)
- PAC-Bayes Learning Bounds for Sample-Dependent Priors (NeurIPS 2020 · similarity 0.89)
- Entropy-SGD optimizes the prior of a PAC-Bayes bound: Generalization properties of Entropy-SGD and data-dependent priors (ICML 2018 · similarity 0.89)
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.