Kosmos: An AI Scientist for Autonomous Discovery
🤖 AI Summary
Current AI research agents lack sufficient coherence to support deep scientific discovery. This paper introduces the first long-horizon coherent AI scientist framework, built upon a multi-agent collaborative architecture that integrates a structured world model, code execution engine, large-scale literature retrieval system, and hypothesis generation module—enabling a closed-loop discovery pipeline: literature reading → hypothesis generation → parallel data analysis → report synthesis. The framework sustains over 200 coherent reasoning steps per session (averaging 42,000 lines of executed code and 1,500 papers read per step), with full provenance for all conclusions. Independent evaluation yields 79.4% accuracy; a single 20-step run equates to six months of human research effort. It has successfully reproduced three previously unpublished findings in metabolomics and materials science, and generated four novel scientific discoveries—significantly advancing the depth, credibility, and automation level of AI-driven scientific research.
Application Category
📝 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.
Problem
Research questions and friction points this paper is trying to address.
Automating iterative scientific discovery cycles of literature search and hypothesis generation
Overcoming AI agent limitations in maintaining coherent long-term research actions
Enhancing traceability and accuracy in automated data-driven scientific reporting
Innovation
Methods, ideas, or system contributions that make the work stand out.
Structured world model enables agent information sharing
Automated cycles of data analysis and literature search
Generates traceable scientific reports with citations
🔎 Similar Papers
No similar papers found.
