HuggingFace Daily Papers·· 1 天前
MLE-World：用学习型世界模型预测机器学习实验结果
MLE-World: Predicting Machine Learning Experiment Outcomes with a Learned World Model
AI 导读
MLE-World 是从 Qwen3.5-4B 微调的世界模型，无需执行代码即可从程序代码和数据描述预测其终端输出（错误轨迹、训练日志、验证分数），并给出预测正确性的置信分数。
正文 · 原文
Abstract:AI research agents can propose, execute, and improve programs that solve machine learning tasks. However, executing each candidate program requires training and evaluating ML models, limiting how many candidates an agent can explore at each step. We introduce MLE-World, a world model finetuned from Qwen3.5-4B that predicts a program's terminal output (error traces, training logs, and validation scores) from its code and data description without executing it, together with a confidence score estimating the prediction's correctness. Because a prediction costs one inference call rather than a training run, MLE-World makes test time scaling (TTS) affordable for these agents. We build five TTS policies, guided by MLE-World, that compare several candidates per step by their predicted outputs: Sequential revision, Best-of-N, Pairwise judging, Fusion, and a UCB style rule that also adds an exploration bonus from this confidence score. Since predictions can be wrong, we run a candidate for real only when its confidence score is low. On 17 MLE-bench Lite tasks, MLE-World matches the real terminal output on 77.9% of held out programs, versus 41.1% for much larger prompted models. MLE-World guided TTS achieves up to 22.1% higher observed mean performance than the agent without a world model, with less real execution. Project webpage: this https URL.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.37175 [cs.LG] |
| (or arXiv:2610.37175v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37175 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Brian Ko [view email]
[v1]
Tue, 27 Oct 2026 13:01:29 UTC (400 KB)
来源：HuggingFace Daily Papers · arxiv.org