Autodata: An agentic data scientist to create high quality synthetic data

[Submitted on 24 Jun 2026 (v1), last revised 4 Jul 2026 (this version, v3)]

Authors:Ilia Kulikov, Chenxi Whitehouse, Tianhao Wu, Yixin Nie, Swarnadeep Saha, Eryk Helenowski, Weizhe Yuan, Olga Golovneva, Jack Lanchantin, Yoram Bachrach, Jakob Foerster, Xian Li, Han Fang, Sainbayar Sukhbaatar, Jason Weston

View PDF HTML (experimental)

Abstract: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.

Submission history

From: Jason Weston [view email]
[v1] Wed, 24 Jun 2026 16:08:31 UTC (19,889 KB)
[v2] Thu, 25 Jun 2026 13:26:50 UTC (19,879 KB)
[v3] Sat, 4 Jul 2026 15:07:51 UTC (19,879 KB)