arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.18378cs.AIcs.CE
收藏

将地球科学模型校准任务委托给知识引导的大语言模型智能体

Delegating Earth science model calibration to knowledge-guided LLM agents

State Key Laboratory of Water Disaster Prevention, Hohai University(河海大学水灾害防御国家重点实验室);Yangtze Institute for Conservation and Development, Hohai University(河海大学长江保护与发展研究院);Meta Platforms Inc.(Meta平台公司);Nanjing Hydraulic Research Institute(南京水利科学研究院)

Zhanliang Zhu, Ziwei Li, Liujun Zhu, Yuchen Liu, Junliang Jin, Jianyun Zhang

AI总结:

该研究开发通用校准工具GCH,将地球科学模型校准委托给知识引导的大语言模型智能体,在雪模型等测试中验证了其高效性与可复现性。

摘要:

将基于过程的地球科学模型应用于新站点和新问题通常需要专家校准决策。可靠地将任务委托给语言模型智能体必须保留科学能力和可复现性。本文开发了一种通用校准工具(GCH),用于从校准请求到评估结果的端到端委托。在知识基础设施(KI)和验证约定的约束下,智能体定义参数域、观测映射和目标,并生成特定于模型的代码。GCH会对照相同的声明检查这些公式,并在数值搜索和评估前将其冻结,此过程无需智能体进一步判断。智能体编写的工作流在六个模型系列上执行,评估识别出未满足的标准。在雪模型测试中,五个知识引导的规范首次尝试执行全部成功,而仅提供模型名称的五个规范执行全部失败,验证纳什-萨特克利夫效率(NSE)达到0.81-0.94,默认值为0.57。在共同目标下,独立的GR4J公式产生的分数在0.001 NSE以内;重新评估十一个保存的参数向量可复现其指标。在十次预算匹配的VIC配对测试中,固定搜索不劣于继续智能体提议(差值为-0.02 NSE;中位数差值为-0.003)。这种分离实现了具有可检查科学选择和可复现数值评估的端到端校准委托。

英文摘要:

Applying process-based Earth science models to new sites and questions often requires expert calibration decisions. Reliable delegation to language-model agents must preserve scientific competence and reproducibility. Here we develop a general calibration harness (GCH) for end-to-end delegation from a calibration request to assessed results. Constrained by a knowledge infrastructure (KI) and a validation convention, agents define parameter domains, observation mappings and objectives, and generate model-specific code. GCH checks the formulations against the same declarations and freezes them before numerical search and assessment, which require no further agent judgement. Agent-authored workflows executed across six model families, with assessment identifying unmet criteria. In snow-model tests, five of five knowledge-guided specifications executed on the first attempt, compared with none of five given only the model name, achieving validation Nash-Sutcliffe efficiency (NSE) of 0.81-0.94 against 0.57 for defaults. Under a common objective, independent GR4J formulations yielded scores within 0.001 NSE; re-evaluating eleven saved parameter vectors reproduced their metrics. In ten paired VIC tests with matched budgets, fixed search was non-inferior to continuing agent proposals (margin, -0.02 NSE; median difference, -0.003). This separation enables end-to-end calibration delegation with inspectable scientific choices and reproducible numerical evaluation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑