Research map / Reinforcement learning / Policy gradient and actor-critic
Control Frequency Adaptation via Action Persistence in Batch 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 |
|---|---|---|---|
| Increase action persistence from 1 to 4 decision steps | PFQI · Cartpole | expected return: 169.9 ± 5.8 -> 239.5 ± 4.4 | Helped |
| Increase action persistence from 1 to 32 decision steps | PFQI · MountainCar | expected return: -111.1 ± 1.5 -> -92.4 ± 1.5 | Helped |
| Increase action persistence from 1 to 8 decision steps | PFQI · LunarLander | expected return: -165.8 ± 50.4 -> 2.0 ± 3.4 | Helped |
| Increase action persistence from 1 to 2 decision steps | PFQI · Pendulum | expected return: -116.7 ± 16.7 -> -113.1 ± 16.3 | No clear effect |
| Increase action persistence from 1 to 2 decision steps | PFQI · Acrobot | expected return: -89.2 ± 1.1 -> -82.5 ± 1.7 | Helped |
| Increase action persistence from 1 to 2 decision steps | PFQI · Swimmer | expected return: 21.3 ± 1.1 -> 25.2 ± 0.8 | Helped |
| Increase action persistence from 1 to 64 decision steps | PFQI · Hopper | expected return: 58.6 ± 4.8 -> 73.4 ± 1.2 | Helped |
| Increase action persistence from 1 to 16 decision steps | PFQI · Walker 2D | expected return: 61.6 ± 5.5 -> 102.1 ± 19.3 | Helped |
| Use higher action persistence instead of k=1 when the batch is small | PFQI · Trading | expected return | Helped |
| Increase batch size | PFQI · Trading | expected return | Helped |
Closest papers by meaning
- Overcoming Slow Decision Frequencies in Continuous Control: Model-Based Sequence Reinforcement Learning for Model-Free Control (ICLR 2025 · similarity 0.87)
- Select before Act: Spatially Decoupled Action Repetition for Continuous Control (ICLR 2025 · similarity 0.87)
- Continuous Q-Score Matching: Diffusion Guided Reinforcement Learning for Continuous-Time Control (NeurIPS 2025 · similarity 0.87)
- Sample Efficient Reinforcement Learning with Partial Dynamics Knowledge (AAAI 2024 · similarity 0.86)
- Sample-Efficiency in Multi-Batch Reinforcement Learning: The Need for Dimension-Dependent Adaptivity (ICLR 2024 · similarity 0.86)
- Exploiting the Replay Memory Before Exploring the Environment: Enhancing Reinforcement Learning Through Empirical MDP Iteration (NeurIPS 2024 · similarity 0.86)
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.