arXiv：cs.LG· Yinling Zhang, Langchen Liu, Dongbin Xiu, Xueyan Zou, Xu Kuang, Mengdi Wang, Shilong Liu·· 4 小时前AI 评分54
SciExam for ENSO 基准测试 AI 智能体能否构建气候模型
SciExam for ENSO: Can AI Agents Build Climate Models?
AI 导读
arXiv 论文（arXiv:2610.10513）提出 SciExam for ENSO 基准，让语言模型智能体在六小时内基于真实观测构建 ENSO 低阶随机气候模型，用冻结的诊断结果作反馈，再由隐藏评分器检验模型对 ENSO 统计特性的复现、未观测变量的恢复和对留出年份的预测。
正文
Abstract:Language-model agents are increasingly asked to carry out open-ended scientific research, yet their results are usually graded against a known answer, a rubric, or a language-model reviewer, none of which can tell whether a new scientific model is valid. The AI Science Exam for El Nino-Southern Oscillation (SciExam for ENSO) is a benchmark in which agents build low-order stochastic models of ENSO, the dominant mode of interannual climate variability, from real observations. Within a six-hour budget, agents process the observations, write their own diagnostics, which are then frozen, and develop a model using only these diagnostics as feedback. Hidden graders then test whether the model reproduces ENSO's statistics, recovers unobserved variables, and forecasts held-out years, and score a published model in the same way. Across twelve agent systems, six produce models that score higher than the published model, mainly through better reconstruction and forecasting. The simplified forms of the stronger models are each compatible with one of the two competing explanations of ENSO's warm-cold asymmetry, an open debate that the task never mentions. Controlled runs of the top system under varied information suggest that its scores do not come from recalling the dated observational record and that the information it receives shapes how it builds its model. SciExam for ENSO can thus evaluate agent research where no answer is known, and the results suggest that agents can already build competitive models whose structures bear on questions that scientists still debate.
| Comments: | 28 pages, 5 figures, 8 tables. Code: this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph) |
| ACM classes: | I.2.6; J.2 |
| Cite as: | arXiv:2610.10513 [cs.AI] |
| (or arXiv:2610.10513v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10513 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yinling Zhang [view email]
[v1]
Wed, 7 Oct 2026 17:52:52 UTC (4,791 KB)
来源：arXiv：cs.LG · arxiv.org