Research map / Reinforcement learning / MDPs and planning
Nearly Minimax Optimal Reinforcement Learning for Linear Markov Decision Processes
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 the proposed algorithm instead of LSVI-UCB | Linear MDP · Linear MDP | regret: $\tilde O\big(\sqrt{d^3H^4K}\big) -> $\tilde O\big(d\sqrt{H^3K}\big) | Helped |
| Use variance-aware weighted ridge regression | Proposed algorithm · Linear MDP | regret: a $\sqrt{H}$ improvement | Helped |
| Estimate variance directly for the optimal value function rather than the optimistic value function | Proposed algorithm · Linear MDP | regret: remove the additional $\sqrt{d}$ dependency | Helped |
| Use rare-switching updates for the value-function estimator | Proposed algorithm · Linear MDP | number of value-function updates: $O(dH \log(1+K/\lambda))$ | Helped |
| Use rare-switching updates rather than updating the value function every episode | Proposed algorithm · Linear MDP | total computational complexity: improves the original LSVI-UCB algorithm by a factor of $K$ | Helped |
Closest papers by meaning
- Achieving Constant Regret in Linear Markov Decision Processes (NeurIPS 2024 · similarity 0.92)
- Provably Efficient RL under Episode-Wise Safety in Constrained MDPs with Linear Function Approximation (NeurIPS 2025 · similarity 0.92)
- Horizon-Free Regret for Linear Markov Decision Processes (ICLR 2024 · similarity 0.90)
- Achieving Tractable Minimax Optimal Regret in Average Reward MDPs (NeurIPS 2024 · similarity 0.90)
- Near-Optimal Regret in Linear MDPs with Aggregate Bandit Feedback (ICML 2024 · similarity 0.90)
- Local Linearity: the Key for No-regret Reinforcement Learning in Continuous MDPs (NeurIPS 2024 · 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.