Research map / Learning theory and algorithms / Sample complexity
Tightening the Dependence on Horizon in the Sample Complexity of Q-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 |
|---|---|---|---|
| sharpened error decomposition and recursion analysis of vanilla synchronous Q-learning | vanilla Q-learning, discounted infinite-horizon MDP · generative model, l-infinity error | sample complexity dependence on 1/(1-γ): from (1-γ)^-5 to (1-γ)^-4 (order-wise improvement of 1/(1-γ)) | Helped |
| rescaled linear or constant learning rates with the new analysis | vanilla Q-learning, finite-horizon MDP · generative model, l-infinity error | sample complexity in H | Helped |
| learning rate schedule choice (rescaled linear vs iteration-invariant constant) | vanilla Q-learning, discounted infinite-horizon MDP · generative model | sample complexity order | No clear effect |
Closest papers by meaning
- Sample Complexity of Asynchronous Q-Learning: Sharper Analysis and Variance Reduction (NeurIPS 2020 · similarity 0.94)
- Sample-Efficient Reinforcement Learning for Linearly-Parameterized MDPs with a Generative Model (NeurIPS 2021 · similarity 0.92)
- Sample Efficient Reinforcement Learning with Partial Dynamics Knowledge (AAAI 2024 · similarity 0.92)
- Optimal Sample Complexity for Average Reward Markov Decision Processes (ICLR 2024 · similarity 0.89)
- Near-Optimal Sample Complexity for MDPs via Anchoring (ICML 2025 · similarity 0.89)
- Span-Based Optimal Sample Complexity for Weakly Communicating and General Average Reward 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.