我们引入了 Autodata, 一种通用方法，使 AI 代理能够充当构建高质量训练和评估数据的数据科学家。我们展示了如何训练 (meta-optimize) 这样的数据科学家代理,，以便它学会创建更强大的数据。我们描述了,的总体表述和具体的实际实现,代理自我指示。我们对计算机科学研究任务,法律推理任务和数学对象推理,进行了实验，与经典的综合数据集创建方法相比，我们获得了改进的结果。对数据科学家代理本身进行进一步,元优化可带来更大的性能提升。代理数据创建提供了一种将增加的推理计算转化为更高质量的模型训练的方法。总体而言,，我们相信这个方向有可能改变我们构建人工智能数据的方式。
We introduce Autodata, a general method that enables AI agents to act as data scientists who build high quality training and evaluation data. We show how to train (meta-optimize) such a data scientist agent, so that it learns to create even stronger data. We describe the overall formulation, and a specific practical implementation, Agentic Self-Instruct. We conduct experiments on computer science research tasks, legal reasoning tasks and reasoning with mathematical objects, where we obtain improved results compared to classical synthetic dataset creation methods. Further, meta-optimizing the data scientist agent itself delivers an even larger performance uplift. Agentic data creation provides a way to convert increased inference compute into higher quality model training. Overall, we believe this direction has the potential to change the way we build AI data.
科目: 人工智能 (cs.AI); 计算和语言 (cs.CL); 机器学习 (cs.LG)
Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)