Research map / Bandits and online learning / Multi-armed bandits
PAC Lower Bounds and Efficient Algorithms for The Max K-Armed Bandit Problem
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 multi-armed adaptive UCB-type sampling (Max-CB) instead of uniform-over-arms unified-arm sampling | Max K-Armed Bandit with few optimal arms · Example 1 synthetic (|K|=10^4) | sample complexity: 3.52e8 vs 6.9e10 for unified-arm algorithm | Helped |
| Use multi-armed Max-CB instead of unified-arm sampling when many arms are near-optimal | Max K-Armed Bandit with many optimal arms · Example 2 synthetic (|K|=10^4) | sample complexity: 1.56e12 vs 6.9e10 for unified-arm algorithm | Hurt |
| Assume lower bound G*(ε) on tail function that is optimistic by factor α<1 (robustness analysis) | Max-CB · theoretical analysis | correctness level and sample complexity | Hurt |
| Use conservative tail bound G*(ε) (true tail is αG*, α>1) | Max-CB · theoretical analysis | failure probability | Helped |
Closest papers by meaning
- PAC Identification of Many Good Arms in Stochastic Multi-Armed Bandits (ICML 2019 · similarity 0.93)
- PAC Identification of a Bandit Arm Relative to a Reward Quantile (AAAI 2017 · similarity 0.91)
- Adaptive Multiple-Arm Identification (ICML 2017 · similarity 0.90)
- Combinatorial Bandits for Maximum Value Reward Function under Value-Index Feedback (ICLR 2024 · similarity 0.90)
- On Interpolating Experts and Multi-Armed Bandits (ICML 2024 · similarity 0.89)
- Oracle-Efficient Combinatorial Semi-Bandits (NeurIPS 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.