Self-Supervised Representation Learning for Astronomical Images
Preprint
Abstract
Sky surveys are the largest data generators in astronomy, making automated tools for extracting meaningful scientific information an absolute necessity. We show that, without the need for labels, self-supervised learning recovers representations of sky survey images that are semantically useful for a variety of scientific tasks. These representations can be directly used as features, or fine-tuned, to outperform supervised methods trained only on labeled data. We apply a contrastive learning framework on multi-band galaxy photometry from the Sloan Digital Sky Survey (SDSS) to learn image representations. We then use them for galaxy morphology classification, and fine-tune them for photometric redshift estimation, using labels from the Galaxy Zoo 2 dataset and SDSS spectroscopy. In both downstream tasks, using the same learned representations, we outperform the supervised state-of-the-art results, and we show that our approach can achieve the accuracy of supervised models while using 2-4 times fewer labels for training.Keywords
All Related Versions
This publication has 32 references indexed in Scilit:
- Effectively using unsupervised machine learning in next generation astronomical surveysAstronomy and Computing, 2021
- A Systematic Review on Supervised and Unsupervised Machine Learning Algorithms for Data Science2019
- Deep learning at scale for the construction of galaxy catalogs in the Dark Energy SurveyPhysics Letters B, 2019
- Photometric redshift estimation via deep learningAstronomy & Astrophysics, 2018
- Measuring photometric redshifts using galaxy images and Deep Neural NetworksAstronomy and Computing, 2016
- Spectroscopic Target Selection in the Sloan Digital Sky Survey: The Main Galaxy SampleThe Astronomical Journal, 2002
- Drizzle: A Method for the Linear Reconstruction of Undersampled ImagesPublications of the Astronomical Society of the Pacific, 2002
- The Sloan Digital Sky Survey Photometric CameraThe Astronomical Journal, 1998
- Slicing Through Multicolor Space: Galaxy Redshifts from Broadband PhotometryThe Astronomical Journal, 1995
- Photometric redshifts of galaxiesThe Astrophysical Journal, 1986