DataMaster: Data-Centric Autonomous AI Research
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916008491483136 |
|---|---|
| author | Du, Yaxin Yang, Xiyuan Zhou, Zhifan Liu, Wanxu Lei, Zixing Chen, Zimeng Liu, Fenyi Wu, Haotian Cai, Yuzhu Liu, Zexi Zhu, Xinyu Wang, WenHao Zhang, Linfeng Qian, Chen Chen, Siheng |
| author_facet | Du, Yaxin Yang, Xiyuan Zhou, Zhifan Liu, Wanxu Lei, Zixing Chen, Zimeng Liu, Fenyi Wu, Haotian Cai, Yuzhu Liu, Zexi Zhu, Xinyu Wang, WenHao Zhang, Linfeng Qian, Chen Chen, Siheng |
| contents | As model families, training recipes, and compute budgets become increasingly standardized, further gains in machine learning systems depend increasingly on data. Yet data engineering remains largely manual and ad hoc: practitioners repeatedly search for external datasets, adapt them to existing pipelines, validate candidate data through downstream training, and carry forward lessons from prior attempts. We study task-conditioned autonomous data engineering, where an autonomous agent improves a fixed learning algorithm by optimizing only the data side, including external data discovery, data selection and composition, cleaning and transformation. The goal is to obtain a stronger downstream solution while leaving the learning algorithm unchanged. To address the open-ended search space, branch-dependent refinement, and delayed validation inherent in autonomous data engineering, we propose DataMaster, a data-agent framework that integrates tree-structured search, shared candidate data, and cumulative memory. DataMaster consists of three key components: a DataTree that organizes alternative data-engineering branches, a shared Data Pool that stores discovered external data sources for reuse, and a Global Memory that records node outcomes, artifacts, and reusable findings. Together, these components allow the agent to discover candidate data, construct executable training inputs, evaluate them through downstream feedback, and carry useful evidence across branches. We evaluate DataMaster on two types of benchmarks, MLE-Bench Lite and PostTrainBench. On MLE-Bench Lite, it improves medal rate by 32.27% over the initial score; on PostTrainBench, it surpasses the instruct model on GPQA (31.02% vs 30.35%). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_10906 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DataMaster: Data-Centric Autonomous AI Research Du, Yaxin Yang, Xiyuan Zhou, Zhifan Liu, Wanxu Lei, Zixing Chen, Zimeng Liu, Fenyi Wu, Haotian Cai, Yuzhu Liu, Zexi Zhu, Xinyu Wang, WenHao Zhang, Linfeng Qian, Chen Chen, Siheng Machine Learning Artificial Intelligence As model families, training recipes, and compute budgets become increasingly standardized, further gains in machine learning systems depend increasingly on data. Yet data engineering remains largely manual and ad hoc: practitioners repeatedly search for external datasets, adapt them to existing pipelines, validate candidate data through downstream training, and carry forward lessons from prior attempts. We study task-conditioned autonomous data engineering, where an autonomous agent improves a fixed learning algorithm by optimizing only the data side, including external data discovery, data selection and composition, cleaning and transformation. The goal is to obtain a stronger downstream solution while leaving the learning algorithm unchanged. To address the open-ended search space, branch-dependent refinement, and delayed validation inherent in autonomous data engineering, we propose DataMaster, a data-agent framework that integrates tree-structured search, shared candidate data, and cumulative memory. DataMaster consists of three key components: a DataTree that organizes alternative data-engineering branches, a shared Data Pool that stores discovered external data sources for reuse, and a Global Memory that records node outcomes, artifacts, and reusable findings. Together, these components allow the agent to discover candidate data, construct executable training inputs, evaluate them through downstream feedback, and carry useful evidence across branches. We evaluate DataMaster on two types of benchmarks, MLE-Bench Lite and PostTrainBench. On MLE-Bench Lite, it improves medal rate by 32.27% over the initial score; on PostTrainBench, it surpasses the instruct model on GPQA (31.02% vs 30.35%). |
| title | DataMaster: Data-Centric Autonomous AI Research |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2605.10906 |