Fudan University · AI³ Institute · Shanghai 复旦大学 · 人工智能创新与产业研究院（AI³）· 上海
原子智能实验室
Building reliable AI scientists to accelerate the discovery of new materials — from atomistic simulation to lab verification. 构建可靠的 AI 科学家，加速新材料研发——从原子模拟到实验验证。
About关于我们
AI for Science & Scientific AI Agents AI for Science 与科学智能体
The Atomistic Intelligence Lab at Fudan University's AI³ Institute designs and builds AI systems that do science. We combine scientific AI agents, atomistic machine learning, quantum chemical simulations, and end-to-end agentic materials research to build reliable AI scientists and accelerate the development of new materials. 复旦大学 AI³ 研究院原子智能实验室（Atomistic Intelligence Lab）致力于 AI for Science 研究。课题组围绕 Scientific AI Agents（科学智能体）、Atomistic Machine Learning（原子体系机器学习）、Quantum Chemical Simulations（量子化学计算）与End-to-end Agentic Materials Research（端到端智能体材料研究）相结合，构建可靠的 AI 科学家，加速新材料的研发。
Scientific AI AgentsScientific AI Agents科学智能体
LLM agents that plan, run, and critique atomistic research.能够规划、执行和评估科学研究的大模型智能体。
Atomistic Machine LearningAtomistic Machine Learning原子体系机器学习
ML for materials, molecules, and drugs: GNNs, MLIPs, and foundation models.面向材料、分子与药物的机器学习：图神经网络、机器学习势函数与基础模型。
Quantum Chemical SimulationsQuantum Chemical Simulations量子化学计算
First-principles simulations (e.g. DFT) to understand and design materials and chemistry.以 DFT 为代表的第一性原理计算，从电子与原子层面理解并设计材料与化学体系。
End-to-end Agentic Materials ResearchEnd-to-end Agentic Materials Research端到端智能体材料研究
Agents that run the full research cycle, from hypothesis to discovery.由智能体端到端完成从提出假设到得出发现的完整科研流程。
Publications论文成果
Selected publications代表性论文
News新闻动态
Latest news最新动态
Join the founding team成为实验室创始成员
We're looking for PhD and Master's students, postdocs, and interns excited about AI for Science.诚招对 AI for Science 感兴趣的博士生、硕士生、博士后与科研实习生。