Research map / Learning theory and algorithms / Sample complexity
POLY-HOOT: Monte-Carlo Planning in Continuous Space MDPs with Non-Asymptotic Analysis
Weichao Mao, Kaiqing Zhang, Qiaomin Xie, Tamer Basar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 POLY-HOOT instead of HOOT | continuous-space MCTS · LunarLander | average reward: -23.37 -> -3.02 | Improved |
| Use POLY-HOOT instead of discretized-UCT | continuous-space MCTS · CartPole-IG | average reward: 69.39 -> 77.85 | Improved |
| Use POLY-HOOT instead of PUCT | continuous-space MCTS · LunarLander | average reward: -43.05 -> -3.02 | Improved |
| Use POLY-HOOT instead of HOOT | continuous-space MCTS · Pendulum | average reward: -109.50 -> -109.43 | No clear effect |
| Use POLY-HOOT instead of HOOT | continuous-space MCTS · CartPole | average reward: 77.85 -> 77.85 | No clear effect |
| Set maximum HOO depth to 2 (ablation) | POLY-HOOT · CartPole-IG | average reward: 77.85 -> 42.45 | Hurt |
| Set maximum HOO depth to 4 (ablation) | POLY-HOOT · CartPole-IG | average reward: 77.85 -> 48.54 | Hurt |
| Set maximum HOO depth to 6 (ablation) | POLY-HOOT · CartPole-IG | average reward: 77.85 -> 63.27 | Hurt |
| Set maximum HOO depth to 10 (ablation) | POLY-HOOT · CartPole-IG | average reward: 77.85 -> 77.85 | No clear effect |
| Limit the HOO tree to depth 8 rather than using HOOT (ablation) | POLY-HOOT · CartPole-IG | time per decision: 1.173 -> 1.030 | Improved |
| Increase maximum HOO depth from 2 to 10 (ablation) | POLY-HOOT · CartPole-IG | time per decision: 0.054 -> 1.057 | Hurt |
Closest papers by meaning
- Power Mean Estimation in Stochastic Continuous Monte-Carlo Tree Search (ICML 2025 · similarity 0.91)
- Graph Sparse Sampling: Breaking the Curse of the Horizon in Continuous MDP Planning (NeurIPS 2026 · similarity 0.87)
- Monte-Carlo Tree Search with Uncertainty Propagation via Optimal Transport (ICML 2025 · similarity 0.86)
- Online Robust Reinforcement Learning Through Monte-Carlo Planning (ICML 2025 · similarity 0.86)
- Graph-Based Stochastic-Power-UCT: Monte-Carlo Graph Search with Power Mean Estimation (NeurIPS 2026 · similarity 0.86)
- Threshold UCT: Cost-Constrained Monte Carlo Tree Search with Pareto Curves (AAAI 2025 · 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.
