AI Paper Atlas

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

MOOSE-Star: Unlocking Tractable Training for Scientific Discovery by Breaking the Complexity Barrier

Zonglin Yang, Lidong Bing

ICML 2026 · Large language models › Agents, code and math · approach: LLMs for scientific discovery · 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
fine-tune on inspiration retrieval data (152k examples)R1-Distilled-Qwen-7B · TOMATO-Star inspiration retrievalaccuracy: 28.42% -> 54.37%Improved
fine-tune on hypothesis composition data (97k samples)R1-Distilled-Qwen-7B · TOMATO-Star hypothesis compositionM3 total: 4.05 -> 4.68Improved
add bounded composition training data (ablation)MS-HC-7B · hypothesis composition with ground-truth inspirationM3 total: 4.68 -> 4.74 (1x), 4.73 (2x)No clear effect
add bounded composition training data (ablation)MS-HC-7B · hypothesis composition with Hard proxy inspirationsM3 total: 2.39 -> 2.42 (1x), 2.56 (2x)Improved
jointly train on IR and HC data (ablation)MS-7B vs MS-IR-7B / MS-HC-7B · TOMATO-StarIR accuracy and HC M3 total: IR 54.37% -> 54.34%; HC 4.68 -> 5.02Improved
use hierarchical best-first search instead of tournament search (ablation)MS-IR-7B routing · 3,035 inspiration corpusIR inference calls: 218.00 -> 67.78Improved
add detailed motivation to hierarchical search (ablation)hierarchical search · 3,035 inspiration corpusIR inference calls: 67.78 -> 63.05Improved
add simple motivation to hierarchical search (ablation)hierarchical search · 3,035 inspiration corpusIR inference calls: 67.78 -> 63.80Improved
increase number of required inspirations k from 1 to 3 (ablation)brute-force end-to-end sampling · per-attempt pass ratepass rate: 9.46% -> 0.23%Hurt
increase number of required inspirations k from 1 to 3 (ablation)decomposed HC and IR subtasks · per-attempt pass ratepass rate: HC 18.88% -> 25.75%; IR 27.84% -> 24.22%No clear effect
scale training samples from 10^0 to 10^5 (ablation)MS-IR-7B · inspiration retrievalaccuracy: 28.4% -> 54.4%Improved
increase inference budget with hierarchical search vs brute-force samplingMOOSE-Star vs brute force · 109 test papersfraction of papers solved: 100% @ 5979 steps vs 38.5% @ 9500Improved

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