Research map / Bandits and online learning / RL theory and MDPs
Minimax Regret Bounds for Reinforcement Learning
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 an empirical-variance Bernstein-Freedman exploration bonus instead of the bounded-value Chernoff-Hoeffding bonus (ablation) | UCBVI · Finite-horizon episodic MDPs | regret upper bound: 20 H^{3/2}L\sqrt{S A K} + 250H^2 S^2 A L^2 -> 30 HL \sqrt{S A K} + 2500H^2 S^2 A L^2+ 4H^{3/2}\sqrt{KL} | Helped |
| Apply concentration inequalities directly to the value function rather than confidence-bounding transition probabilities | UCBVI-CH · Finite-horizon episodic MDPs | regret upper bound: \widetilde{O}(HS \sqrt{AT}) -> \widetilde O(H\sqrt{SAT}) | Helped |
| Use empirical next-state value variance in the exploration bonus and couple the per-step variances through the law of total variance (ablation) | UCBVI-BF · Finite-horizon episodic MDPs | regret upper bound: \widetilde O(H\sqrt{SAT}) -> \widetilde O(\sqrt{HSAT}) | Helped |
| Use UCBVI-BF in the large-T regime | UCBVI-BF · Finite-horizon episodic MDPs, T\geq H^3S^3A and SA\geq H | regret upper bound: \widetilde O(\sqrt{HSAT}) | Helped |
Closest papers by meaning
- Regret Minimization for Reinforcement Learning by Evaluating the Optimal Bias Function (NeurIPS 2019 · similarity 0.92)
- First-Order Regret in Reinforcement Learning with Linear Function Approximation: A Robust Estimation Approach (ICML 2022 · similarity 0.92)
- Horizon-Free Regret for Linear Markov Decision Processes (ICLR 2024 · similarity 0.91)
- Minimax Optimal Regret Bound for Reinforcement Learning with Trajectory Feedback (ICML 2025 · similarity 0.91)
- Efficient Exploration in Average-Reward Constrained Reinforcement Learning: Achieving Near-Optimal Regret With Posterior Sampling (ICML 2024 · similarity 0.90)
- Position: Constants are Critical in Regret Bounds for Reinforcement Learning (ICML 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.