Research map / Bandits and online learning / Multi-armed bandits
Anytime optimal algorithms in stochastic multi-armed bandits
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 MOSS-anytime (MOSS-style index with no doubling trick) instead of MOSS with a doubling trick | stochastic bandit with equal gaps · synthetic Gaussian bandit, equal gaps | regret | Helped |
| Use MOSS-anytime instead of UCB2 and other anytime algorithms | stochastic bandit with increasing gaps · synthetic Gaussian bandit, increasing gaps | regret | Hurt |
| Set exploration parameter alpha to 0.1 instead of theoretical 1.35 (ablation) | MOSS-anytime · increasing gaps synthetic bandit | regret | Helped |
| Use a single-pull variant of UCB2 instead of block-structured UCB2 (ablation) | UCB2 · synthetic Gaussian bandit | regret | No clear effect |
| Add the conjectured-optimal H-inspired UCB index (sum of pulls below sqrt t) | UCB variant · equal gaps and increasing gaps synthetic bandits | regret | No clear effect |
| Use full-information follow-the-leader (FTL) instead of bandit feedback | full information setting · theory | distribution-free regret bound | Helped |
Closest papers by meaning
- Anytime Exploration for Multi-armed Bandits using Confidence Information (ICML 2016 · similarity 0.91)
- Stochastic Multi-Armed Bandits with Unrestricted Delay Distributions (ICML 2021 · similarity 0.91)
- Stochastic Rising Bandits (ICML 2022 · similarity 0.90)
- Probably Anytime-Safe Stochastic Combinatorial Semi-Bandits (ICML 2023 · similarity 0.90)
- Combinatorial Stochastic-Greedy Bandit (AAAI 2024 · similarity 0.88)
- Leveraging (Biased) Information: Multi-armed Bandits with Offline Data (ICML 2024 · 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.