AI Paper Atlas

Research map / Large language models / Agents, code and math

Subgoal Search For Complex Reasoning Tasks

Konrad Czechowski, Tomasz Odrzygóźdź, Marek Zbysiński, Michał Zawalski, Krzysztof Olejnik, Yuhuai Wu, Łukasz Kuciński, Piotr Miłoś

NeurIPS 2021 · Large language models › Agents, code and math · approach: Inductive and abstract reasoning evaluation · 0 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 subgoal generator (BF-kSubS) instead of low-level-action BestFSbest-first search · INT proof length 15, graph size 50success rate: 0.09 -> 0.38Improved
use subgoal generator instead of low-level actions at larger budgetbest-first search · INT proof length 15, graph size 400success rate: 0.47 -> 0.91Improved
use subgoal generator on Sokoban 20x20best-first search · Sokoban 20x20, graph size 50success rate: 0.02 -> 0.46Improved
use subgoal generator on Sokoban 12x12 large budgetbest-first search · Sokoban 12x12, graph size 5000success rate: 0.83 -> 0.93Improved
use subgoal generator on Rubik's Cubebest-first search · Rubik's Cubesuccess rate: less than 10% -> nearly 100%Improved
use MCTS with subgoals instead of AlphaZeroMCTS · INT proof length 15success rateImproved
increase subgoal distance k (ablation)BF-kSubS · INT and Rubik's Cubesuccess rateMixed
increase subgoal distance k up to 10 (ablation)BF-kSubS · Sokobansuccess rateImproved
use subgoals with k=4 instead of single steps on noisy value function (ablation)best-first search · Grid world sigma=10success rate: 0.142 -> 1Improved
use subgoals with k=4 under highest value noise (ablation)best-first search · Grid world sigma=20success rate: 0.006 -> 0.983Improved
use subgoals with k=4 under low value noise (ablation)best-first search · Grid world sigma=3success rate: 0.999 -> 1No clear effect
move k=4 steps ahead instead of 1 step along solution path (ablation)value estimates · INTprobability value decreases: 0.32 -> 0.02Improved

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