[论文解读] Kosmos: An AI Scientist for Autonomous Discovery

Ludovico Mitchener, Angela Yiu+35 more|ArXiv.org|Nov 4, 2025

Biomedical Text Mining and Ontologies被引用 3

一句话总结

Kosmos 通过一个结构化的世界模型，协调数据分析与文献检索，在最多12小时循环内实现数据驱动的科学发现、可追溯的科学报告以及跨领域发现，并显著节省专家时间。

ABSTRACT

Data-driven scientific discovery requires iterative cycles of literature search, hypothesis generation, and data analysis. Substantial progress has been made towards AI agents that can automate scientific research, but all such agents remain limited in the number of actions they can take before losing coherence, thus limiting the depth of their findings. Here we present Kosmos, an AI scientist that automates data-driven discovery. Given an open-ended objective and a dataset, Kosmos runs for up to 12 hours performing cycles of parallel data analysis, literature search, and hypothesis generation before synthesizing discoveries into scientific reports. Unlike prior systems, Kosmos uses a structured world model to share information between a data analysis agent and a literature search agent. The world model enables Kosmos to coherently pursue the specified objective over 200 agent rollouts, collectively executing an average of 42,000 lines of code and reading 1,500 papers per run. Kosmos cites all statements in its reports with code or primary literature, ensuring its reasoning is traceable. Independent scientists found 79.4% of statements in Kosmos reports to be accurate, and collaborators reported that a single 20-cycle Kosmos run performed the equivalent of 6 months of their own research time on average. Furthermore, collaborators reported that the number of valuable scientific findings generated scales linearly with Kosmos cycles (tested up to 20 cycles). We highlight seven discoveries made by Kosmos that span metabolomics, materials science, neuroscience, and statistical genetics. Three discoveries independently reproduce findings from preprinted or unpublished manuscripts that were not accessed by Kosmos at runtime, while four make novel contributions to the scientific literature.

研究动机与目标

• 通过迭代的数据分析与文献综述循环，推动自治科学发现。
• 利用共享的世界模型实现对开放性目标的协同多智能体探索。
• 产出具可科学报告性且可追溯数据与文献引证的发现。

提出的方法

• 在并行运行的两名 Edison Sci 代理（数据分析与文献检索）之间协同工作。
• 维护结构化的世界模型，在代理之间共享与综合输出。
• 每个循环执行最多十项任务并迭代更新世界模型。
• 生成三到四份科学报告，报告中的陈述与数据或文献紧密相关。
• 通过对陈述、文献来源和分析的专家评估来评估报告的准确性。
• 将探索扩展到最多200个代理回滚，单次运行约编写约42,000行代码。

Figure 1: Kosmos workflow and performance. a) Overall workflow for Kosmos. (left) Kosmos is provided with an initial dataset and broad research objective specified by a scientist. (middle) The Kosmos world model coordinates data analysis and literature search agents to identify key discoveries. (rig

实验结果

研究问题

• RQ1 一个以 AI 为驱动的系统是否可以自动进行迭代数据分析与文献综述，以完成开放式研究目标？
• RQ2 结构化世界模型是否能够实现跨代理连贯推理并为可追溯的科学报告引用来源？
• RQ3 在不同领域，Kosmos 生成的发现的准确性与专家时间等效性如何？

主要发现

• Kosmos 在每次运行中大约执行 42,000 行代码，阅读约 1,500 篇论文。
• 独立专家评估认为 Kosmos 的 79.4% 的陈述是准确的。
• 协作者估算，20轮 Kosmos 运算相当于大约六个月的人力研究时间。
• Kosmos 产生了横跨代谢组学、材料科学、神经科学和统计遗传学的七项发现。
• 三项发现复现了未发表或截止后发现；四项贡献了新见解。
• Kosmos 报告中的所有陈述都链接到数据分析输出或可追溯的文献引用。

Figure 2: Kosmos reproduces unpublished discovery in neuroprotection metabolomics. a) Research objective and dataset description provided to Kosmos and experimental design for untargeted metabolomic profiling of mouse brains. b) Barplot showing up-regulation of purine and pyrimidine salvage products