arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.09031cs.LGphysics.ao-phstat.AP
收藏

生活在地球观测的物理学和机器学习相互作用中

Living in the Physics and Machine Learning Interplay for Earth Observation

Image Processing Laboratory (IPL), Universitat de València(瓦伦西亚大学图像处理实验室 (IPL))

Gustau Camps-Valls, Daniel H. Svendsen, Jordi Cortés-Andrés, Álvaro Moreno-Martínez, Adrián Pérez-Suay, Jose Adsuara, Irene Martín, Maria Piles, Jordi Muñoz-Marí, Luca Martino

摘要:

地球科学中的大多数问题旨在对系统进行推断,而准确的预测只是整个问题的一小部分。推理意味着理解变量之间的关系,推导出物理上可解释的、简单的、数学上可处理的模型。机器学习模型本身就是优秀的近似器,但通常不尊重最基本的物理定律,如质量或能量守恒,因此一致性和置信度受到损害。在本文中,我们描述了该领域面临的主要挑战,并介绍了物理学和机器学习相互作用的几种方式:从数据中编码微分方程,用物理先验和依赖约束约束数据驱动的模型,改进参数化,仿真物理模型,以及混合数据驱动和基于过程的模型。这是一个集体的长期人工智能议程,旨在开发和应用能够在地球系统中发现知识的算法。

英文摘要:

Most problems in Earth sciences aim to do inferences about the system, where accurate predictions are just a tiny part of the whole problem. Inferences mean understanding variables relations, deriving models that are physically interpretable, that are simple parsimonious, and mathematically tractable. Machine learning models alone are excellent approximators, but very often do not respect the most elementary laws of physics, like mass or energy conservation, so consistency and confidence are compromised. In this paper, we describe the main challenges ahead in the field, and introduce several ways to live in the Physics and machine learning interplay: to encode differential equations from data, constrain data-driven models with physics-priors and dependence constraints, improve parameterizations, emulate physical models, and blend data-driven and process-based models. This is a collective long-term AI agenda towards developing and applying algorithms capable of discovering knowledge in the Earth system.

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。

↑