arXiv cs.MA· Geonhee Ahn, Donghyun Lee, Hayoung Doo, Jonggeol Na, Hyunsoo Cho, Sookyung Kim·· 11 小时前AI 评分29
MIND：面向材料研究的 AI 联合科学家框架
MIND: AI Co-Scientist for Material Research
AI 导读
MIND 是一个 LLM 驱动的材料研究框架，将科学发现流程组织为假设精炼、实验和基于辩论的验证多智能体流水线。系统集成机器学习原子间势（尤其是 SevenNet-Omni）实现可扩展的硅基实验，并提供网页界面用于自动化假设测试。模块化设计支持接入更多实验模块，代码已开源，论文被 ECML PKDD 2026 Demonstration Track 接收。
正文 · 原文
Abstract:Large language models (LLMs) have enabled agentic AI systems for scientific discovery, but most approaches remain limited to textbased reasoning without automated experimental verification. We propose MIND, an LLM-driven framework for automated hypothesis validation in materials research. MIND organizes the scientific discovery process into hypothesis refinement, experimentation, and debate-based validation within a multi-agent pipeline. For experimental verification, the system integrates Machine Learning Interatomic Potentials, particularly SevenNet-Omni, enabling scalable in-silico experiments. We also provide a web-based user interface for automated hypothesis testing. The modular design allows additional experimental modules to be integrated, making the framework adaptable to broader scientific workflows. The code is available at: this https URL, and a demonstration video at: this https URL.
| Comments: | 4 pages, 3 figures. Accepted at ECML PKDD 2026, Demonstration Track. Code: this https URL . Demo video: this https URL |
| Subjects: | Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE) |
| Cite as: | arXiv:2604.13699 [cs.MA] |
| (or arXiv:2604.13699v2 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2604.13699 arXiv-issued DOI via DataCite |
|
| Journal reference: | In: Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track, Demo Track and Industrial Track, LNCS 16950, pp. 324-328 (2027) |
| Related DOI: | https://doi.org/10.1007/978-3-032-37685-5_27
DOI(s) linking to related resources |
Submission history
From: Dong Hyun Lee [view email]
[v1]
Wed, 15 Apr 2026 10:27:01 UTC (2,028 KB)
[v2]
Wed, 4 Nov 2026 20:55:55 UTC (2,028 KB)
来源：arXiv cs.MA · arxiv.org