Research map / Learning theory and algorithms / Clustering
Improving PAC Exploration Using the Median Of Means
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 |
|---|---|---|---|
| replace simple average over samples with median of means in Bellman backups | Median-PAC · 5x5 gridworld | consistency of finding optimal policy | Helped |
| bounds scale with variance of Bellman operator instead of Qmax^2 (median of means) | Median-PAC · theoretical PAC bounds (discrete MDPs) | sample complexity dependence: trades a factor of Qmax^2 for sigma^2; improves MORMAX by 1/(1-gamma)^2 even in worst case | Helped |
| replace average with median of means in Pazis and Parr delayed-update/continuous algorithm | Pazis and Parr [12] algorithm · continuous state, concurrent, delayed update settings (theoretical) | PAC bound: improves bounds by (Rmax + gamma Qmax)^2 while introducing a factor of sigma^2 | Helped |
Closest papers by meaning
- Span-Based Optimal Sample Complexity for Weakly Communicating and General Average Reward MDPs (NeurIPS 2024 · similarity 0.84)
- Optimal Non-Asymptotic Rates of Value Iteration for Average-Reward Markov Decision Processes (ICLR 2025 · similarity 0.84)
- Sharp Gap-Dependent Variance-Aware Regret Bounds for Tabular MDPs (NeurIPS 2025 · similarity 0.83)
- Adaptive Exploration for Multi-Reward Multi-Policy Evaluation (ICML 2025 · similarity 0.83)
- A Unified Recipe for Deriving (Time-Uniform) PAC-Bayes Bounds (NeurIPS 2024 · similarity 0.83)
- Finite-Time Bounds for Average-Reward Fitted Q-Iteration (NeurIPS 2025 · similarity 0.83)
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.