Research map / Reinforcement learning / MDPs and planning
Tight Regret Bounds for Model-Based Reinforcement Learning with Greedy Policies
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 full-planning with greedy policies using 1-step planning | UCRL2 · Finite-horizon tabular MDPs | regret: UCRL2: \tilde O(\sqrt{H^2S^2AT}) -> UCRL2-GP: \Olog\br*{HS\sqrt{AT} +H^2\sqrt{S}SA} | No clear effect |
| Replace full-planning with greedy policies using 1-step planning | EULER · Finite-horizon tabular MDPs | regret: EULER: \Olog(\sqrt{HSAT}) -> EULER-GP: \Olog\br*{\sqrt{HS A T} + \sqrt{S}SAH^2(\sqrt{S}+\sqrt{H})} | No clear effect |
| Replace full-planning with greedy policies using 1-step planning | UCRL2 · Finite-horizon tabular MDPs | time complexity per episode: \Olog(\mathcal{N} SAH) -> \Olog(\mathcal{N} AH) | Helped |
| Replace full-planning with greedy policies using 1-step planning | EULER · Finite-horizon tabular MDPs | time complexity per episode: \mathcal{O}(\mathcal{N} S A H) -> \mathcal{O}(\mathcal{N} A H) | Helped |
| Use greedy policy updates instead of full-planning | UCRL2 and EULER · Chain environment and 2D chain | performance | No clear effect |
Closest papers by meaning
- Model-based RL as a Minimalist Approach to Horizon-Free and Second-Order Bounds (ICLR 2025 · similarity 0.92)
- Horizon-Free Regret for Linear Markov Decision Processes (ICLR 2024 · similarity 0.90)
- Efficient Exploration in Average-Reward Constrained Reinforcement Learning: Achieving Near-Optimal Regret With Posterior Sampling (ICML 2024 · similarity 0.90)
- Hybrid Reinforcement Learning Breaks Sample Size Barriers In Linear MDPs (NeurIPS 2024 · similarity 0.89)
- DARLING: Detection Augmented Reinforcement Learning with Non-Stationary Guarantees (NeurIPS 2026 · similarity 0.89)
- Minimax Optimal Regret Bound for Reinforcement Learning with Trajectory Feedback (ICML 2025 · similarity 0.89)
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.