Research map / Probabilistic and scientific modeling / Molecules and proteins
Monte Carlo Tree Search With Iteratively Refining State Abstractions
Samuel Sokota, Caleb Ho, Zaheen Farraz Ahmad, J Zico Kolter
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 |
|---|---|---|---|
| Replace progressive widening with abstraction refining | MCTS state selector · Toy Markov reward process with heterogeneous stochasticity | value error | Improved |
| Set progressive widening alpha to a small value (ablation) | Progressive widening · Toy MRP | number of level-two children | Hurt |
| Set progressive widening alpha to a large value (ablation) | Progressive widening · Toy MRP | focus on level-two children | Hurt |
| Use abstraction refining with a single hyperparameter setting instead of tuned progressive widening | Policy evaluation · Continuous blackjack | value error vs iterations | Improved |
| Use small alpha in progressive widening (ablation) | Progressive widening · Continuous blackjack | value error stability | Hurt |
| Use abstraction refining instead of progressive widening | MCTS control · Trap problem | expected return | Improved |
| Use a distance function that cannot distinguish falling into the trap (ablation) | Abstraction refining · Trap problem | expected return | Hurt |
| Use abstraction refining instead of progressive widening (k=1) | AlphaZero latent-state search · Five by five Go opponent exploitation | territory margin | Mixed |
| Increase rollout budget and alpha (ablation) | Progressive widening · Five by five Go | expected return | Improved |
Closest papers by meaning
- Monte Carlo Tree Search in the Presence of Transition Uncertainty (AAAI 2024 · similarity 0.90)
- Monte Carlo Tree Diffusion for System 2 Planning (ICML 2025 · similarity 0.89)
- Graph-Based Stochastic-Power-UCT: Monte-Carlo Graph Search with Power Mean Estimation (NeurIPS 2026 · similarity 0.89)
- Online Robust Reinforcement Learning Through Monte-Carlo Planning (ICML 2025 · similarity 0.88)
- Monte-Carlo Tree Search with Uncertainty Propagation via Optimal Transport (ICML 2025 · similarity 0.88)
- Threshold UCT: Cost-Constrained Monte Carlo Tree Search with Pareto Curves (AAAI 2025 · similarity 0.88)
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.
