Grounding LLMs in Scientific Discovery via Embodied Actions

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Bo, Zhou, Jinfeng, Chen, Yuxuan, Yin, Jianing, Huang, Minlie, Wang, Hongning
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917291187240960
author Zhang, Bo
Zhou, Jinfeng
Chen, Yuxuan
Yin, Jianing
Huang, Minlie
Wang, Hongning
author_facet Zhang, Bo
Zhou, Jinfeng
Chen, Yuxuan
Yin, Jianing
Huang, Minlie
Wang, Hongning
contents Large Language Models (LLMs) have shown significant potential in scientific discovery but struggle to bridge the gap between theoretical reasoning and verifiable physical simulation. Existing solutions operate in a passive "execute-then-response" loop and thus lacks runtime perception, obscuring agents to transient anomalies (e.g., numerical instability or diverging oscillations). To address this limitation, we propose EmbodiedAct, a framework that transforms established scientific software into active embodied agents by grounding LLMs in embodied actions with a tight perception-execution loop. We instantiate EmbodiedAct within MATLAB and evaluate it on complex engineering design and scientific modeling tasks. Extensive experiments show that EmbodiedAct significantly outperforms existing baselines, achieving SOTA performance by ensuring satisfactory reliability and stability in long-horizon simulations and enhanced accuracy in scientific modeling.
format Preprint
id arxiv_https___arxiv_org_abs_2602_20639
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Grounding LLMs in Scientific Discovery via Embodied Actions
Zhang, Bo
Zhou, Jinfeng
Chen, Yuxuan
Yin, Jianing
Huang, Minlie
Wang, Hongning
Artificial Intelligence
Large Language Models (LLMs) have shown significant potential in scientific discovery but struggle to bridge the gap between theoretical reasoning and verifiable physical simulation. Existing solutions operate in a passive "execute-then-response" loop and thus lacks runtime perception, obscuring agents to transient anomalies (e.g., numerical instability or diverging oscillations). To address this limitation, we propose EmbodiedAct, a framework that transforms established scientific software into active embodied agents by grounding LLMs in embodied actions with a tight perception-execution loop. We instantiate EmbodiedAct within MATLAB and evaluate it on complex engineering design and scientific modeling tasks. Extensive experiments show that EmbodiedAct significantly outperforms existing baselines, achieving SOTA performance by ensuring satisfactory reliability and stability in long-horizon simulations and enhanced accuracy in scientific modeling.
title Grounding LLMs in Scientific Discovery via Embodied Actions
topic Artificial Intelligence
url https://arxiv.org/abs/2602.20639