Research map / Reinforcement learning / MDPs and planning
Efficient Reinforcement Learning in Factored MDPs with Application to Constrained RL
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 factorized empirical-variance confidence bonuses and refined variance decomposition | FMDP-BF · Episodic FMDP | regret upper bound: by a factor of \sqrt{nH\Gamma} | Helped |
| Apply the refined variance decomposition theorem | FMDP-BF compared with concurrent UCBVI and EULER methods · Episodic FMDP | regret upper bound: by a factor of $\sqrt{n}$ | Helped |
| Represent remaining budget as factored state and estimate the cost distribution | FMDP-BF for RLwK · RL with knapsack constraints | regret upper bound: O\left(\sqrt{HSA(1+dBm)T}\right) to roughly $O\left(\sqrt{HdSAT}\right)$ | Helped |
| Use Hoeffding-type instead of empirical-variance confidence bonuses | FMDP-CH relative to FMDP-BF · Episodic FMDP | regret upper bound | Hurt |
Closest papers by meaning
- Overcoming the Curse of Dimensionality in Reinforcement Learning Through Approximate Factorization (ICML 2025 · similarity 0.92)
- Provably Efficient RL under Episode-Wise Safety in Constrained MDPs with Linear Function Approximation (NeurIPS 2025 · similarity 0.88)
- Efficient Exploration in Average-Reward Constrained Reinforcement Learning: Achieving Near-Optimal Regret With Posterior Sampling (ICML 2024 · similarity 0.88)
- Confident Natural Policy Gradient for Local Planning in $q_\pi$-realizable Constrained MDPs (NeurIPS 2024 · similarity 0.87)
- Near-Optimal Sample Complexity for Online Constrained MDPs (NeurIPS 2025 · similarity 0.87)
- A Primal-Dual Algorithm for Offline Constrained Reinforcement Learning with Linear MDPs (ICML 2024 · similarity 0.87)
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.