王赤,李晖,罗冰显,沈芳,王晶晶,张灵倩,杨易,赵栋.人工智能赋能空间科学——以空间天气为例[J].中国科学院院刊,2026,41(6):1169-1181.

人工智能赋能空间科学——以空间天气为例

Artificial intelligence empowering space science—Case study of space weather

王赤^1,2*
中国科学院国家空间科学中心 北京 100190;中国科学院大学 地球与行星科学学院 北京 100049
WANG Chi^1,2*
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China;College of Earth and Planetary Sciences, University of Chinese Academy of Sciences, Beijing 100049, China
李晖^1,3
中国科学院国家空间科学中心 北京 100190;中国科学院大学 天文与空间科学学院 北京 100049
LI Hui^1,3
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China;School of Astronomy and Space Science, University of Chinese Academy of Sciences, Beijing 100049, China
罗冰显^1,3
中国科学院国家空间科学中心 北京 100190;中国科学院大学 天文与空间科学学院 北京 100049
LUO Bingxian^1,3
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China;School of Astronomy and Space Science, University of Chinese Academy of Sciences, Beijing 100049, China
沈芳^1,2
中国科学院国家空间科学中心 北京 100190;中国科学院大学 地球与行星科学学院 北京 100049
SHEN Fang^1,2
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China;College of Earth and Planetary Sciences, University of Chinese Academy of Sciences, Beijing 100049, China
王晶晶¹
中国科学院国家空间科学中心 北京 100190
WANG Jingjing¹
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China
张灵倩¹
中国科学院国家空间科学中心 北京 100190
ZHANG Lingqian¹
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China
杨易¹
中国科学院国家空间科学中心 北京 100190
YANG Yi¹
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China
赵栋¹
中国科学院国家空间科学中心 北京 100190
ZHAO Dong¹
National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China

         空间科学;人工智能;研究范式;空间天气

        space science;artificial intelligence;research paradigm;space weather

        空间科学研究正面临海量观测数据爆发式增长、物理过程跨尺度强耦合，以及国家战略需求日益迫切的三重挑战。传统研究范式在分析时效、预测精度与自主化水平上的短板，难以有效保障在轨卫星安全、重大航天任务顺利实施等重大需求。文章构建空间科学智能化“感知—认知—决策”3层架构，以空间天气这一强应用属性的领域为典型案例，系统剖析人工智能在物理约束建模、可解释因果推断、自主智能体应用及星地协同观测等关键方向带来的四维范式变革实践，深入探讨智能化转型中面临的深层次科学与技术挑战，并提出面向空间科学全域发展的战略举措，旨在为我国在全球空间科学智能化竞争中抢占先机、实现“领跑”提供可落地的技术路径与政策框架。

        Space science is currently confronted with a triple challenge: the explosive growth of observational data, the strongly coupled cross-scale nature of physical processes, and the increasingly urgent national strategic demands. The limitations of traditional research paradigms in analytical efficiency, forecast accuracy, and autonomous capability hinder their effectiveness in meeting critical requirements such as safeguarding on-orbit satellites and ensuring the successful execution of major space missions. This study proposes a three-layer “perception–cognition–decision-making” architecture for intelligent space science. Taking space weather—a domain with strong operational relevance—as a representative case, the four-dimensional paradigm transformation driven by artificial intelligence is systematically examined across key directions, including physics-constrained modeling, explainable causal inference, autonomous agent applications, and satellite – ground collaborative observation. The profound scientific and technical challenges confronting the intelligent transformation and strategic initiatives are further explored for the holistic advancement of space science. This work aims to provide actionable technical pathways and policy frameworks for China to seize the initiative and achieve leadership in the global race of intelligent space science.