研究人员开发了一种新颖的方法来优化机器学习气候模型的训练数据集，增强其泛化到新场景的能力。通过使用可微分的简单气候模型（SCM），他们迭代地调整训练数据以最大化模拟器的技能。这种方法在单个优化场景上进行训练，其性能优于在多个标准ScenarioMIP路径上训练的模拟器，用更少的数据实现了更高的预测精度。优化后的训练数据使模拟器能够更好地分离温室气体和气溶胶等不同气候强迫因子的独特物理行为。 AI
影响 优化的训练数据生成可能导致更高效、更准确的气候模拟器，降低气候建模的计算成本。
排序理由 该集群包含一篇详细介绍机器学习气候模型新方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2606.19302v1 Announce Type: cross Abstract: As deep learning for physical systems continues to grow in popularity, efforts to improve generalizability have primarily focused on designing architectures that embed physical constraints. However, for machine-learning surrogate …
As deep learning for physical systems continues to grow in popularity, efforts to improve generalizability have primarily focused on designing architectures that embed physical constraints. However, for machine-learning surrogate climate models (emulators), we show that the low s…