标题： AI宇宙学家 I：用于自动化数据分析的能动系统 显示英文标题

标题： The AI Cosmologist I: An Agentic System for Automated Data Analysis

摘要： 我们介绍了AI宇宙学家，这是一个设计用于自动化宇宙学/天文学数据分析和机器学习研究工作流程的主动系统。 该系统从创意生成到实验评估和研究成果传播，实现了完整的管道，模仿了通常由人类研究人员执行的科学过程。 该系统采用专门的规划、编码、执行、分析和综合代理协同工作，以开发新的方法。 与传统的自动机器学习系统不同，AI宇宙学家生成多样化的实现策略，编写完整代码，处理执行错误，分析结果，并根据实验结果综合出新方法。 我们在多个机器学习任务中展示了AI宇宙学家的能力，展示了它如何能够成功探索解决方案空间，基于实验结果迭代，并结合不同方法的成功元素。 我们的结果显示，主动系统可以自动化部分研究过程，可能加速科学发现。 本文使用的代码和实验数据可在GitHub上获取，网址为https://github.com/adammoss/aicosmologist。 附录中的示例论文展示了系统从仅有的数据集和任务描述开始，自主生成完整科学出版物的能力。 显示英文摘要

摘要： We present the AI Cosmologist, an agentic system designed to automate cosmological/astronomical data analysis and machine learning research workflows. This implements a complete pipeline from idea generation to experimental evaluation and research dissemination, mimicking the scientific process typically performed by human researchers. The system employs specialized agents for planning, coding, execution, analysis, and synthesis that work together to develop novel approaches. Unlike traditional auto machine-learning systems, the AI Cosmologist generates diverse implementation strategies, writes complete code, handles execution errors, analyzes results, and synthesizes new approaches based on experimental outcomes. We demonstrate the AI Cosmologist capabilities across several machine learning tasks, showing how it can successfully explore solution spaces, iterate based on experimental results, and combine successful elements from different approaches. Our results indicate that agentic systems can automate portions of the research process, potentially accelerating scientific discovery. The code and experimental data used in this paper are available on GitHub at https://github.com/adammoss/aicosmologist. Example papers included in the appendix demonstrate the system's capability to autonomously produce complete scientific publications, starting from only the dataset and task description