Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Phased Exploration with Greedy Exploitation in Stochastic Combinatorial Partial Monitoring Games
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 GCB's confidence-bound exploration with phased forced exploration plus greedy exploitation (PEGE) | CPM learner · Stochastic CPM regret | distribution independent regret | Helped |
| PEGE with C(a)=h a, alpha=1, beta=1 (linearly growing exploration, exponential exploitation) | PEGE · Stochastic CPM regret | distribution dependent regret | Helped |
| Add gap estimation stage before PEGE (PEGE2), with fallback to pure exploitation when threshold T0 exceeded | PEGE · Stochastic CPM regret | distribution dependent regret | Helped |
| Improper tuning of h in PEGE can add a large additive constant to regret; h must scale with the gap | PEGE · Stochastic CPM regret | additive regret constant | Hurt |
| Online ranking with feedback at top: PEGE applied when unique optimal ranking assumption fails | PEGE · Online ranking with top feedback (stochastic) | regret order | No clear effect |
Closest papers by meaning
- Instance-Dependent Regret Bounds for Nonstochastic Linear Partial Monitoring (NeurIPS 2025 · similarity 0.88)
- Bayesian Analysis of Combinatorial Gaussian Process Bandits (ICLR 2025 · similarity 0.87)
- Oracle-Efficient Combinatorial Semi-Bandits (NeurIPS 2025 · similarity 0.87)
- Infrequent Exploration in Linear Bandits (NeurIPS 2025 · similarity 0.86)
- Random Exploration in Bayesian Optimization: Order-Optimal Regret and Computational Efficiency (ICML 2024 · similarity 0.86)
- Randomized Confidence Bounds for Stochastic Partial Monitoring (ICML 2024 · 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.