Solaris: A Foundation Model for the Sun

Abstract

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. Our experiments show that increasing the number of model parameters leads to improved performance, aligning with established scaling laws in other domains. 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.

Please cite this paper as:

@inproceedings{majid2024solaris,
  title={Solaris: A Foundation Model for the Sun},
  author={Majid, Harris Abdul and Sittoni, Pietro and Tudisco, Francesco},
  booktitle={NeurIPS 2024 Workshop on Foundation Models for Science},
  year={2024}
}

Keywords: Deep learning neural networks foundation models generative AI AI4Science