arXiv cs.AI· Shangheng Du, Jinxin Shi, Shuaiyu Zhang, Yiming Wu, Junpeng Zhao, Shiyang Feng, Xin Li, Liang He, Xiangchao Yan, Bo Zhang·· 2 天前AI 评分60
MLExpert：面向自主机器学习工程的通用智能体模型
MLExpert: Towards a Generalist Agentic Model for Autonomous Machine Learning Engineering
AI 导读
研究提出面向自主长周期机器学习工程优化的后训练方法，并据此训练出 35B MoE 智能体模型 MLExpert。该方法通过 Kaggle-MLE Factory 构建可执行任务与交互环境，用分层智能体 SFT 迁移工具使用和编码能力，并用 Windowed Interaction RL 把相邻实验结果之间的多轮决策作为一个执行窗口赋予共享结果奖励。
正文 · 原文
Abstract:Large language models (LLMs) are expanding from assisting scientific research and engineering to autonomously developing and improving AI systems, advancing AI for AI (AI4AI) as a pathway toward recursive self-improvement (RSI). Workflow-driven machine learning engineering (MLE) systems provide a practical path toward this pursuit through complex orchestration that manages long-horizon experimentation, while recent MLE systems adopt model-directed interaction, allowing models to autonomously organize experiments and adapt optimization strategies based on execution feedback. However, feedback from each experiment reflects several preceding decisions, and a long MLE run contains many such costly experiments. Training must therefore select useful supervision from long interaction trajectories and learn from intermediate outcomes. We propose a post-training method for autonomous long-horizon MLE optimization and apply it to train MLExpert, a 35B Mixture-of-Experts agentic model. We design the Kaggle-MLE Factory to construct executable tasks and an interaction harness for model-directed experimentation. Hierarchical agentic SFT transfers general tool-use and coding capabilities to MLE through outcome-guided supervision. Windowed Interaction RL (WIRL) treats the multi-turn decisions between successive experiment results as an execution window and assigns it a shared outcome-based reward. Asynchronous collection and policy correction support policy updates before the trajectory ends. On MLE-Bench Lite, MLExpert improves the average medal rate of the base model from 37.9% to 72.7%, achieving performance comparable to larger frontier models such as Kimi-K3 and GLM-5.3. Further experiments show that MLExpert improves performance on NatureBench Lite, agent scaffold optimization, and algorithm optimization, suggesting that the learned optimization capability generalizes to other tasks.
| Comments: | Code: this https URL Model: this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.22344 [cs.AI] |
| (or arXiv:2610.22344v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22344 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Shangheng Du [view email]
[v1]
Thu, 15 Oct 2026 17:56:32 UTC (764 KB)
来源：arXiv cs.AI · arxiv.org