Research map / Bandits and online learning / Multi-armed bandits
Adaptive Algorithms for Infinitely Many-Armed Bandits: A Unified Framework
Emmanuel Pilliat
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| re-rank arms by LCB and restrict the exploration scope to high-LCB arms (OSE -> PROSE) | OSE · Beta(1,1/alpha) arms, alpha in {0.25,0.5,1,2} | simple regret | Improved |
| use deterministic exponentially growing exploration scopes with LCB ranking (PROSE) instead of BSH doubling-trick brackets | BSH · Runtime, K=5000, Tmax=50000 | mean time per trial (ms): 89 -> 76 | Improved |
| use the LCB-ranking modification of PROSE instead of OSE under small beta (ablation) | OSE · Beta(1,1/alpha) arms, alpha in {0.5,1}, beta=1 | stability of simple regret | Improved |
| increase beta from small values to 50 (ablation) | OSE · Beta(1,1/alpha) arms, alpha in {0.5,1} | simple regret | Hurt |
| PROSE with beta=50 compared with BSH (ablation) | PROSE · Beta(1,1/alpha) arms, alpha in {0.5,1} | simple regret | Improved |
| add LCB ranking (PROSE) instead of OSE or BSH for cumulative regret | OSE · Beta(1,1/alpha) arms, t much larger than K | cumulative regret trend | Improved |
| naive UCB (scope infinite) instead of OSE or BSH | OSE · Beta(1,1/alpha) arms, t below K | simple regret | Hurt |
| OSE compared with BSH | OSE · Beta(1,1/alpha) arms, alpha in {0.25,0.5,1,2} | simple regret | Mixed |
Closest papers by meaning
- Infinitely Many-Armed Bandits with Budget Constraints (AAAI 2017 · similarity 0.93)
- Adaptation to the Range in K-Armed Bandits (JMLR 2023 · similarity 0.93)
- From Finite to Countable-Armed Bandits (NeurIPS 2020 · similarity 0.92)
- Adaptive Regret for Bandits Made Possible: Two Queries Suffice (ICLR 2024 · similarity 0.92)
- An Adaptive Approach for Infinitely Many-armed Bandits under Generalized Rotting Constraints (NeurIPS 2024 · similarity 0.91)
- Greedy Algorithms for Structured Bandits: A Sharp Characterization of Asymptotic Success / Failure (NeurIPS 2025 · similarity 0.90)
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.
