Research map / Reinforcement learning / MDPs and planning
Incremental Stochastic Factorization for Online 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 EMSF's data-distribution-weighted KL instead of uniform-row KL (KLM NMF) when sampling is heterogeneous | EMSF · Random SF transition matrices, 90% mass on 50% of states | approximation error | Helped |
| use alpha < 1 in incremental EMSF (ablation) | Incremental EMSF · Random transition matrix (n=100, m=30) | KL convergence speed | Helped |
| use larger commit interval tc together with larger alpha (ablation) | Incremental EMSF · Random transition matrix (n=100, m=30) | KL convergence | Mixed |
| EMSF+PISF instead of CNT+PI using a model 5% the size of the MDP | EMSF+PISF · Blackjack | return | Helped |
| model-based methods instead of Q-learning | EMSF+PISF · Blackjack | return | Helped |
| EMSF+PISF instead of CNT+PI | EMSF+PISF · Blackjack | execution time: 16.1 s -> 59.6 s | Hurt |
Closest papers by meaning
- Overcoming the Curse of Dimensionality in Reinforcement Learning Through Approximate Factorization (ICML 2025 · similarity 0.88)
- Efficient Reinforcement Learning in Factored MDPs with Application to Constrained RL (ICLR 2021 · similarity 0.88)
- Randomized Exploration for Reinforcement Learning with Multinomial Logistic Function Approximation (NeurIPS 2024 · similarity 0.86)
- Exploiting the Replay Memory Before Exploring the Environment: Enhancing Reinforcement Learning Through Empirical MDP Iteration (NeurIPS 2024 · similarity 0.85)
- Deterministic Uncertainty Propagation for Improved Model-Based Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.85)
- Logarithmic Regret for Online KL-Regularized Reinforcement Learning (ICML 2025 · similarity 0.85)
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.