Research map / Reinforcement learning / Policy gradient and actor-critic
Estimating the Maximum Expected Value in Continuous Reinforcement Learning Problems
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 Weighted FQI-inf (continuous actions) instead of FQI with 100 discretized actions | FQI with GP · Swing-up pendulum continuous actions, 5 episodes | average reward: 0.412 -> 0.426 | Helped |
| use Weighted FQI-inf instead of FQI with discretized actions | FQI with GP · Swing-up pendulum continuous actions, 20 episodes | average reward: 0.793 -> 0.823 | Helped |
| use Weighted FQI-inf instead of Double FQI | Double FQI · Swing-up pendulum continuous actions, 5 episodes | average reward: 0.120 -> 0.426 | Helped |
| use Weighted estimator with GP (infinite actions) instead of ME/DE with binned actions | WE-inf · Pricing problem | mean bias | Helped |
| increase number of bins for ME and DE (ablation) | ME, DE · Pricing problem | mean bias | Hurt |
| use Weighted FQI (finite actions) instead of FQI | FQI with GP · Swing-up pendulum discrete actions | average reward | Helped |
| use Double FQI instead of FQI | FQI with GP · Swing-up pendulum discrete actions, few samples | average reward | Hurt |
| use WE-inf instead of ME/DE on variance of estimate | WE-inf · Pricing problem | variance of bias | No clear effect |
Closest papers by meaning
- Smoothed Action Value Functions for Learning Gaussian Policies (ICML 2018 · similarity 0.88)
- Policy Evaluation for Variance in Average Reward Reinforcement Learning (ICML 2024 · similarity 0.87)
- Oracle-Efficient Reinforcement Learning for Max Value Ensembles (NeurIPS 2024 · similarity 0.86)
- Stochastic Q-learning for Large Discrete Action Spaces (ICML 2024 · similarity 0.86)
- Dynamic Uncertainty Estimation for Offline Reinforcement Learning (AAAI 2025 · similarity 0.86)
- Enhancing Value Function Estimation through First-Order State-Action Dynamics in Offline Reinforcement Learning (ICML 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.