Research map / Bandits and online learning / Multi-armed bandits
Bootstrapping Upper Confidence Bound
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 multiplier-bootstrap quantiles with a second-order correction instead of concentration-based confidence bounds | UCB for truncated-normal multi-armed bandits · Truncated-normal bandit | cumulative regret | Helped |
| Use multiplier-bootstrap confidence bounds | UCB for Gaussian multi-armed bandits · Gaussian bandit | cumulative regret | Mixed |
| Use multiplier-bootstrap confidence bounds instead of concentration-based UCB | UCB for Bernoulli and Beta bandits · Bernoulli and Beta bandits | cumulative regret | Helped |
| Use multiplier-bootstrap confidence bounds instead of Bernoulli Thompson sampling | UCB for Beta bandits · Beta bandit | cumulative regret | Helped |
| Use multiplier-bootstrap confidence bounds instead of Bernoulli Thompson sampling | UCB for Bernoulli bandits · Bernoulli bandit | cumulative regret | Hurt |
| Use bootstrapped confidence bounds rather than Bernoulli-TS and Vanilla UCB under noise-standard-deviation misspecification | Bootstrapped UCB · Gaussian bandit, cumulative regret at T=2000 | cumulative regret | Helped |
| Extend the bootstrap confidence-bound approach to linear bandits | Bootstrapped UCB for linear bandits · Linear bandit | cumulative regret | Helped |
| Remove the second-order correction from the bootstrapped threshold (ablation) | Bootstrapped UCB · Two-arm Bernoulli bandit | expected cumulative regret: R(T)\geq \Delta_2\Big((1-\mu_1)\mu_2(T-2)+1\Big) | Hurt |
| Use the naive bootstrap without second-order correction for small samples (ablation) | Bootstrap confidence bound for the sample mean · Truncated-normal mean estimation, sample size \le 10 | 95% confidence-bound validity | Hurt |
| Add the second-order correction to the bootstrap quantile (ablation) | Bootstrap confidence bound for the sample mean · Truncated-normal mean estimation | confidence-bound sharpness | Helped |
Closest papers by meaning
- Gaussian Process Upper Confidence Bound Achieves Nearly-Optimal Regret in Noise-Free Gaussian Process Bandits (NeurIPS 2025 · similarity 0.89)
- Improved Online Confidence Bounds for Multinomial Logistic Bandits (ICML 2025 · similarity 0.88)
- Noise-Adaptive Confidence Sets for Linear Bandits and Application to Bayesian Optimization (ICML 2024 · similarity 0.87)
- Bayesian Regret Minimization in Offline Bandits (ICML 2024 · similarity 0.87)
- Leveraging (Biased) Information: Multi-armed Bandits with Offline Data (ICML 2024 · similarity 0.87)
- Improved Regret Bounds for Gaussian Process Upper Confidence Bound in Bayesian Optimization (NeurIPS 2025 · 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.