首个太阳大气预测基础模型，用13年多波段数据训练。
Solaris: A Foundation Model of the Sun
- 基于10900万参数的3D Swin Transformer架构，预训练12小时预报。
- 仅用1700 Å单波段微调即超越从零训练模型。
- 适合太阳物理研究者，推动空间天气预测发展。
基础模型已在多个科学领域展现卓越性能，激发我们探索其在太阳物理学中的潜力。本文提出Solaris，首个用于预测太阳大气的基础模型。我们利用太阳动力学观测台（SDO）连续13年、覆盖完整太阳周期的全盘多波段图像数据，对Solaris进行12小时间隔预报的预训练。Solaris采用大规模3D Swin Transformer架构，参数量达1.09亿。通过在低数据条件下对未参与预训练的1700 Å单波段进行微调，Solaris表现出优异泛化能力，优于从零开始训练的模型。结果表明，Solaris能有效捕捉太阳大气的复杂动态，显著提升太阳预报水平。
原文摘要 · Abstract (English)
Foundation models have demonstrated remarkable success across various scientific domains, motivating our exploration of their potential in solar physics. In this paper, we present Solaris, the first foundation model for forecasting the Sun's atmosphere. We leverage 13 years of full-disk, multi-wavelength solar imagery from the Solar Dynamics Observatory, spanning a complete solar cycle, to pre-train Solaris for 12-hour interval forecasting. Solaris is built on a large-scale 3D Swin Transformer architecture with 109 million parameters. We demonstrate Solaris' ability to generalize by fine-tuning on a low-data regime using a single wavelength (1700 Å), that was not included in pre-training, outperforming models trained from scratch on this specific wavelength. Our results indicate that Solaris can effectively capture the complex dynamics of the solar atmosphere and transform solar forecasting.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。