发表机构
University of Toronto; Acceleration Consortium; Vector Institute for Artificial Intelligence; Canadian Institute for Advanced Research (CIFAR); NVIDIA; University of Amsterdam; Instituto de Micro y Nanotecnología, IMN-CNM, CSIC; Sungkyunkwan University; Merck KGaA（多伦多大学; 加速联盟; 向量人工智能研究所; 加拿大高级研究院; 英伟达; 阿姆斯特丹大学; 西班牙国家研究委员会微纳米技术研究所; 成均馆大学; 默克集团）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出La Agente Óptima智能体框架，可监督贝叶斯优化实验方案，经多任务评估，其能维持实验方案、提升产率、降低成本，拓展了自动驾驶实验室的应用范围。
AI 中文摘要
自动驾驶实验室（SDLs）将自动化实验与自适应决策相结合，以加速科学发现，然而其运行往往依赖人类专家，专家需将科学目标转化为可执行的闭环实验方案，并根据数据和操作条件的变化对方案进行调整。在此，我们提出La Agente Óptima，这是一种智能体框架，可在计算系统和实验系统中构建并监督贝叶斯优化实验方案，同时维持持久的优化状态。通过将大语言模型（LLM）推理与已执行的实验方案分离，Óptima能够持续运行重复的优化循环，仅在进展需要解释或方案修订时将控制权交还给智能体，并确保每一项决策都可审计。我们通过消融研究、5项数字发现任务和2个物理平台对Óptima进行评估。在所有测试中，Óptima均能在科学问题和执行环境发生变化时维持可执行的实验方案。在一项闭环接触角优化实验中，Óptima识别并纠正了运行中途的测量故障，将接触角从71.4度提升至67.8度，略高于64-66度的目标范围；基于此结果，Óptima正确推断出使用现有试剂可能无法达到目标，并建议更改配方。在一项为期5天的多目标流动化学实验中，Óptima通过23次实验将产率从30%提升至59%；尽管存在可观的推理成本，其成本低于人类主导的实验，且使用的起始材料显著更少，同时选择了质量效率更高的操作点。这些结果表明，基于LLM的智能体可让领域科学家无需专家设置即可开展严谨、长期的优化实验，拓展了SDLs的应用范围。
英文摘要
Self-driving laboratories (SDLs) combine automated experimentation with adaptive decision-making to accelerate scientific discovery. Their operation nevertheless often depends on human specialists who translate scientific objectives into executable closed-loop campaigns. Specialists adjust them as data and operating conditions change. Here, we present La Agente Óptima, an agentic framework that constructs and supervises Bayesian optimization campaigns across computational and experimental systems while maintaining a persistent optimization state. By separating large language model (LLM) reasoning from executed campaigns, Óptima runs repetitive optimization loops consistently, returns control to the agent only when progress requires interpretation or campaign revision, and keeps every decision auditable. We evaluate Óptima across ablation studies, five digital discovery tasks, and two physical platforms. Throughout, Óptima maintained executable campaigns as both the scientific problem and execution environment evolved. In a closed-loop contact angle optimization campaign, Óptima identified and corrected a mid-run measurement failure, bringing the contact angle from 71.4 to 67.8 degrees, just above the 64-66 degree range. From this result, Óptima correctly inferred that the target was likely unattainable with the available reagents and recommended changing the formulation. In a five-day multi-objective flow-chemistry campaign, Óptima increased the yield from 30% to 59% over 23 experiments. Despite substantial inference costs, it cost less and used substantially less starting material than a human-directed campaign, while selecting a more mass-efficient operating point. These results show that LLM-based agents can make rigorous, long-running optimization campaigns accessible to domain scientists without specialist setup, expanding the scope of SDLs.
