Research map / Learning theory and algorithms / Combinatorial and search optimization
Grouping Nodes with known Value Differences: A lossless UCT-based Abstraction Algorithm
Robin Schmöcker, Alexander Dockhorn, Bodo Rosenhahn
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 |
|---|---|---|---|
| group state-action pairs by known value differences (KVDA) instead of equal value | OGA-UCT · deterministic environments, 1000 iterations | average return: -1255.6 -> -1158.2 (d-Manufacturer) | Improved |
| group by known value differences (KVDA) | OGA-UCT · d-Tamarisk, 1000 iterations | average return: -214.5 -> -185.6 | Improved |
| group by known value differences (KVDA) | OGA-UCT · d-Othello, 1000 iterations | average return: 181.2 -> 179.6 | No clear effect |
| group by known value differences (KVDA) | OGA-UCT · abstraction detection, 1000 iterations | ratio of non-trivial abstract Q nodes: 0.48 -> 0.15 (d-SysAdmin) | Improved |
| group by known value differences (KVDA) | OGA-UCT · d-Constrictor abstraction detection | ratio of non-trivial abstract Q nodes: 0.97 -> 0.98 | No clear effect |
| ignore rewards entirely when abstracting ((inf,0)-OGA) (ablation) | OGA-UCT · d-Earth Observation, 1000 iterations | average return: -7.18 -> -30.0 | Hurt |
| ignore rewards entirely when abstracting ((inf,0)-OGA) (ablation) | OGA-UCT · d-Push Your Luck, 1000 iterations | average return: 125.1 -> 66.7 | Hurt |
| use KVDA with single parameter C=4 instead of OGA variants | OGA-UCT and (eps_a,0)-OGA · normalized pairings score, deterministic environments | normalized pairings score | Improved |
| replace KVDA with eps_t-KVDA in stochastic environments | (eps_a,eps_t)-OGA · stochastic environments | average return | No clear effect |
| replace KVDA with eps_t-KVDA in stochastic environments | (eps_a,eps_t)-OGA · s-Tamarisk | average return | Hurt |
| add KVDA to OGA-UCT (runtime) | OGA-UCT · decision time | decision time: median overhead < 1% (100 iterations) | No clear effect |
Closest papers by meaning
- Monte Carlo Tree Search in the Presence of Transition Uncertainty (AAAI 2024 · similarity 0.88)
- Graph-Based Stochastic-Power-UCT: Monte-Carlo Graph Search with Power Mean Estimation (NeurIPS 2026 · similarity 0.88)
- Monte-Carlo Tree Search with Uncertainty Propagation via Optimal Transport (ICML 2025 · similarity 0.87)
- Improving Monte Carlo Tree Search for Symbolic Regression (NeurIPS 2025 · similarity 0.86)
- An Efficient Algorithm for Thresholding Monte Carlo Tree Search (NeurIPS 2026 · similarity 0.85)
- Beyond Self-Repellent Kernels: History-Driven Target Towards Efficient Nonlinear MCMC on General Graphs (ICML 2025 · similarity 0.85)
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.
