Research map / Bandits and online learning / Multi-armed bandits
Identifying Outlier Arms in Multi-Armed Bandit
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 |
|---|---|---|---|
| Replace round-robin pulling with weighted round-robin (active arms pulled ρ times as often) | WRR vs RR · Synthetic | #pulls: WRR saves more than 40% of RR cost in 80% of test cases; >60% in about half | Helped |
| Replace round-robin pulling with weighted round-robin | WRR vs RR · Twitter | cost reduction vs IB: WRR saves >50% of IB cost in ~80% of cases; RR often worse than IB | Helped |
| Use a heuristic termination (stop when outlier set unchanged for 1/δ pulls) in naive round-robin | NRR · Synthetic | % exactly correct | Hurt |
| Iterative best-arm identification baseline vs. RR/WRR | IB · Synthetic | #pulls: 1-3 orders of magnitude more pulls than WRR and RR | Hurt |
| Set ρ to ρ* from Eq. (6) rather than a manually tuned value (ablation) | WRR · Synthetic (n=15, k=2.5) | avg #pulls ratio to ρ*: ρ* performance close to best (ratio within about 1.08 of ρ*) | No clear effect |
Closest papers by meaning
- Adaptive Double-Exploration Tradeoff for Outlier Detection (AAAI 2020 · similarity 0.91)
- Robust Outlier Arm Identification (ICML 2020 · similarity 0.90)
- MaxGap Bandit: Adaptive Algorithms for Approximate Ranking (NeurIPS 2019 · similarity 0.89)
- Balancing Performance and Costs in Best Arm Identification (NeurIPS 2025 · similarity 0.87)
- Robustly Improving Bandit Algorithms with Confounded and Selection Biased Offline Data: A Causal Approach (AAAI 2024 · similarity 0.86)
- The Batch Complexity of Bandit Pure Exploration (ICML 2025 · similarity 0.86)
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.