用大模型主动设计实验，闭环优化科学发现过程。
LLM-AutoSciLab: Closed-Loop Scientific Discovery via Active Experimentation with LLMs
- 大模型自动生成假设并选择能区分假设的实验
- 在化学酶动力学和基因调控网络任务中准确率超基线
- 实验效率提升2-5倍，适合科研自动化场景
科学发现是假设引导数据采集、观测结果反向修正假设的闭环过程。现有方法多依赖固定数据集的监督学习，难以应对局部拟合但无法泛化的多重机制。核心挑战在于如何选择具有信息量的观测以减少不确定性，需从静态推断转向动态数据获取。为此，我们提出LLM-AutoSciLab，一个将假设生成与条件实验选择、机制优化结合的闭环框架。不被动拟合数据，而是迭代生成合理假设，选择可区分或细化假设的实验，并用新证据更新认知。为评估动态闭环发现中的主动数据获取，我们构建ActiveSciBench，包含57个酶动力学任务（ActiveSciBench-Chem）和45个基因调控网络任务（ActiveSciBench-GRN），模拟预算受限下的自适应实验设计、变量选择与真机制恢复。在NewtonBench、ActiveSciBench-Chem和ActiveSciBench-GRN上，该框架分别取得67.6%、35.1%符号准确率及31.1%精确图恢复率；且假设引导实验比最强基线节省2-5倍样本。
原文摘要 · Abstract (English)
Scientific discovery is a closed-loop process in which hypotheses guide data acquisition and observations refine the hypothesis space. Yet most approaches reduce discovery to supervised learning over fixed datasets, where limited observations can support multiple plausible mechanisms that fit locally but fail to generalize. Thus, the key challenge is selecting informative observations to resolve uncertainty, shifting the focus from static inference to adaptive data acquisition. To address this, we propose LLM-AutoSciLab, a closed-loop framework that couples hypothesis generation with hypothesis-conditioned experiment selection and mechanism refinement. Rather than fitting models to passively collected data, LLM-AutoSciLab iteratively proposes plausible hypotheses, selects informative experiments to distinguish or refine them, and updates its state using the resulting evidence. To evaluate dynamic, closed-loop scientific discovery with active data acquisition, we introduce ActiveSciBench, comprising two datasets: ActiveSciBench-Chem with 57 enzyme-kinetics tasks and ActiveSciBench-GRN with 45 gene-regulatory-network tasks. These datasets model discovery as a budget-constrained process requiring adaptive experiment design, variable selection, and recovery of true mechanisms. Across NewtonBench, ActiveSciBench-Chem, and ActiveSciBench-GRN, LLM-AutoSciLab outperforms prior methods, achieving 67.6% and 35.1% symbolic accuracy on NewtonBench and ActiveSciBench-Chem, respectively, and 31.1% exact graph recovery on ActiveSciBench-GRN. Moreover, hypothesis-guided experimentation is 2-5x more sample-efficient than the strongest competing baselines. Code and data are available at: https://github.com/scientific-discovery/LLM-AutoSciLab
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。