AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

Spending Thinking Time Wisely: Accelerating MCTS with Virtual Expansions

Weirui Ye, Pieter Abbeel, Yang Gao

NeurIPS 2022 · Learning theory and algorithms › Combinatorial and search optimization · approach: Monte Carlo tree search · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
Use adaptive virtual-expansion termination instead of a fixed 150-iteration searchEfficientZero MCTS · Go 9×9 vs GNU Go level 10average budget: 150 ± 0.0 -> 76 ± 10.8Improved
Use adaptive virtual-expansion termination instead of a fixed 150-iteration searchEfficientZero MCTS · Go 9×9 vs GNU Go level 10winning rate: 75% ± 3.0% -> 71% ± 4.7%Hurt
Use adaptive virtual-expansion termination instead of 50 fixed iterationsEfficientZero MCTS · Pongscore: 19.7 ± 1.6 -> 18.8 ± 2.8No clear effect
Use adaptive virtual-expansion termination instead of 50 fixed iterationsEfficientZero MCTS · Breakoutscore: 410.7 ± 15.1 -> 372.8 ± 18.3Hurt
Use adaptive virtual-expansion termination instead of 50 fixed iterationsEfficientZero MCTS · Seaquestscore: 1159.9 ± 90.7 -> 970.0 ± 339.5Hurt
Use adaptive virtual-expansion termination instead of 50 fixed iterationsEfficientZero MCTS · Heroscore: 9992.1 ± 2059.4 -> 8928.1 ± 2922.1Hurt
Use adaptive virtual-expansion termination instead of 50 fixed iterationsEfficientZero MCTS · Qbertscore: 14495.8 ± 683.9 -> 11476.6 ± 978.2Hurt
Add virtual expansion rather than stopping at 30 vanilla iterations (ablation)V-MCTS · Go 9×9winning rate: 17% ± 3.2% -> 32% ± 3.5%Improved
Add virtual expansion rather than greedy expansion for the remaining simulations (ablation)V-MCTS · Go 9×9winning rate: 3% ± 2.0% -> 32% ± 3.5%Improved
Replace the L1 termination criterion with L2 at ε=0.1 (ablation)V-MCTS · Go 9×9winning rate: 81.5% -> 79.8%Hurt
Replace the L1 termination criterion with L2 at ε=0.1 (ablation)V-MCTS · Go 9×9average budget: 96.2 -> 97.1Hurt
Use L2 termination and reduce ε from 0.1 to 0.05 (ablation)V-MCTS · Go 9×9winning rate: 79.8% -> 81.0%Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy