Research map / Reinforcement learning / MDPs and planning
Reward-Mixing MDPs with Few Latent Contexts are Learnable
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 higher-order moment matching without requiring recovery of latent reward parameters | EM^2 for general RMMDPs · General RMMDPs | episodes: O(\epsilon^{-2}) instead of~O(\epsilon^{-4}) | Helped |
| Apply EM^2 to the balanced two-context case, using first- and second-order moments to determine third-order moments | Balanced 2-RMMDPs · Balanced 2-RMMDPs | episodes: O(poly(H,Z)(SA)^2/\epsilon^4) -> O(poly(H,Z)(SA)^2/\epsilon^2) | Helped |
| Restrict reward probabilities to a finite integral grid | EM^2 · RMMDPs with integral reward probabilities | episodes: (SA)^{O(M)} -> (SA)^{O(\log M)} | Helped |
| Use bandit feedback with a moment-matching hard instance | General RMMDPs · General RMMDPs | episodes: (SA)^{\Omega(\sqrt{M})}/\epsilon^2 | Hurt |
Closest papers by meaning
- Reinforcement Learning in Reward-Mixing MDPs (NeurIPS 2021 · similarity 0.96)
- Towards Minimax Optimal Reward-free Reinforcement Learning in Linear MDPs (ICLR 2023 · similarity 0.91)
- Optimal Sample Complexity for Average Reward Markov Decision Processes (ICLR 2024 · similarity 0.89)
- Global Convergence for Average Reward Constrained MDPs with Primal-Dual Actor Critic Algorithm (NeurIPS 2025 · similarity 0.88)
- Optimal Single-Policy Sample Complexity and Transient Coverage for Average-Reward Offline RL (NeurIPS 2025 · similarity 0.88)
- Towards Global Optimality for Practical Average Reward Reinforcement Learning without Mixing Time Oracles (ICML 2024 · similarity 0.88)
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.