SpectraFM

Self-supervised learning model excels at analyzing auroral spectra

Researchers have developed a self-supervised learning approach for analyzing auroral emission spectra, utilizing a 1D Vision Transformer pre-trained on a large dataset of unlabelled spectra. This model effectively recov…