2000 character limit reached
Rare Galaxy Classes Identified In Foundation Model Representations (2312.02910v1)
Published 5 Dec 2023 in astro-ph.GA and cs.CV
Abstract: We identify rare and visually distinctive galaxy populations by searching for structure within the learned representations of pretrained models. We show that these representations arrange galaxies by appearance in patterns beyond those needed to predict the pretraining labels. We design a clustering approach to isolate specific local patterns, revealing groups of galaxies with rare and scientifically-interesting morphologies.
- Galaxy zoo: Bulgeless galaxies with growing black holes. Monthly Notices of the Royal Astronomical Society, 429(3):2199–2211, 3 2013. ISSN 13652966. doi: 10.1093/mnras/sts491. URL https://academic.oup.com/mnras/article-lookup/doi/10.1093/mnras/sts491.
- R. Grouchy and R. Buta. A photometric and spectroscopic study of non-barred ringed galaxies. Proceedings of the International Astronomical Union, 2(S235):100, 8 2006. ISSN 17439213. doi: 10.1017/S1743921306005357.
- No assembly required: Mergers are mostly irrelevant for the growth of low-mass dwarf galaxies. Monthly Notices of the Royal Astronomical Society, 479(1):319–331, 9 2018. ISSN 13652966. doi: 10.1093/mnras/sty1488. URL https://academic.oup.com/mnras/article/479/1/319/5033708.
- The formation and evolution of low-surface-brightness galaxies. Monthly Notices of the Royal Astronomical Society, 485(1):796–818, 2 2019. ISSN 13652966. doi: 10.1093/mnras/stz356. URL http://arxiv.org/abs/1902.04580.
- Jellyfish: The origin and distribution of extreme ram-pressure stripping events in massive galaxy clusters. Monthly Notices of the Royal Astronomical Society, 455(3):2994–3008, 1 2016. ISSN 13652966. doi: 10.1093/mnras/stv2508. URL https://academic.oup.com/mnras/article/455/3/2994/2892617.
- Locations and Morphologies of Jellyfish Galaxies in A2744 and A370. The Astrophysical Journal, 937(1):18, 9 2022. ISSN 0004-637X. doi: 10.3847/1538-4357/AC8B6E. URL https://iopscience.iop.org/article/10.3847/1538-4357/ac8b6ehttps://iopscience.iop.org/article/10.3847/1538-4357/ac8b6e/meta.
- Zoobot: Adaptable Deep Learning Models for Galaxy Morphology. Journal of Open Source Software, 8(85):5312, 2023a. doi: 10.21105/joss.05312. URL https://doi.org/10.21105/joss.05312.
- Karen L. Masters. Twelve years of Galaxy Zoo. In Proceedings of the International Astronomical Union, volume 14, pages 205–212, 10 2019. doi: 10.1017/S1743921319008615. URL http://arxiv.org/abs/1910.08177.
- Practical galaxy morphology tools from deep supervised representation learning. Monthly Notices of the Royal Astronomical Society, 513(2):1581–1599, 5 2022a. ISSN 13652966. doi: 10.1093/mnras/stac525. URL https://academic.oup.com/mnras/article/513/2/1581/6539338.
- Classification and Anomaly Detection for Astronomical Survey Data. In Springer Series in Astrostatistics, volume 1, pages 149–184. Springer, New York, NY, 2013. ISBN 9781461435082. doi: 10.1007/978-1-4614-3508-2{_}8. URL https://link.springer.com/chapter/10.1007/978-1-4614-3508-2_8.
- Systematic serendipity: a test of unsupervised machine learning as a method for anomaly detection. Monthly Notices of the Royal Astronomical Society, 484(1):834–849, 3 2019. ISSN 0035-8711. doi: 10.1093/mnras/sty3461. URL https://academic.oup.com/mnras/article/484/1/834/5256651.
- Anomaly detection in Hyper Suprime-Cam galaxy images with generative adversarial networks. Monthly Notices of the Royal Astronomical Society, 508(2):2946–2963, 5 2021. ISSN 13652966. doi: 10.1093/mnras/stab2589. URL http://arxiv.org/abs/2105.02434.
- Allan Sandage. Classification and Stellar Content of Galaxies Obtained from Direct Photography. Galaxies and the Universe. Edited by Allan Sandage, Mary Sandage, and Jerome Kristian, with an index prepared by Gustav A. Tammann. Published by the University of Chicago Press (Stars and Stellar Systems. Volume 9), Chicago, IL USA., 1975, p.1, page 1, 1 1975. URL https://ui.adsabs.harvard.edu/abs/1975gaun.book....1S/abstract.
- An objective classification system for spiral galaxies. I. The two dominant dimensions. ApJ, 278:61–80, 3 1984. ISSN 0004-637X. doi: 10.1086/161768. URL https://ui.adsabs.harvard.edu/abs/1984ApJ...278...61W/abstract.
- An automatic taxonomy of galaxy morphology using unsupervised machine learning. Monthly Notices of the Royal Astronomical Society, 473(1):1108–1129, 2018. ISSN 13652966. doi: 10.1093/mnras/stx2351.
- Beyond the hubble sequence - Exploring galaxy morphology with unsupervised machine learning. Monthly Notices of the Royal Astronomical Society, 503(3):4446, 2021. ISSN 13652966. doi: 10.1093/mnras/stab734. URL http://arxiv.org/abs/2009.11932.
- EfficientNet: Rethinking model scaling for convolutional neural networks. In 36th International Conference on Machine Learning, ICML 2019, pages 10691–10700, Long Beach, CA, 5 2019. Proceedings of Machine Learning Research. ISBN 9781510886988. URL http://arxiv.org/abs/1905.11946.
- Ross Wightman. PyTorch Image Models. https://github.com/rwightman/pytorch-image-models, 2019.
- MaxViT: Multi-Axis Vision Transformer. In European Conference of Computer Vision (ECCV), 4 2022. doi: 10.48550/arxiv.2204.01697. URL https://arxiv.org/abs/2204.01697v4.
- Overview of the DESI Legacy Imaging Surveys. The Astronomical Journal, 157(5):168, 4 2019. ISSN 00046256. doi: 10.3847/1538-3881/ab089d. URL http://arxiv.org/abs/1804.08657.
- Galaxy Zoo DESI: Detailed Morphology Measurements for 8.7M Galaxies in the DESI Legacy Imaging Surveys. MNRAS, 000:1–18, 9 2023b. URL https://arxiv.org/abs/2309.11425v1.
- UMAP: Uniform Manifold Approximation and Projection. Journal of Open Source Software, 3(29):861, 2 2018. ISSN 2475-9066. doi: 10.21105/joss.00861. URL https://doi.org/10.21105/joss.00861.
- hdbscan: Hierarchical density based clustering. The Journal of Open Source Software, 2(11):205, 2017. ISSN 2475-9066. doi: 10.21105/joss.00205.
- Density-preserving data visualization unveils dynamic patterns of single-cell transcriptomic variability. bioRxiv, 2020. doi: 10.1101/2020.05.12.077776. URL https://www.biorxiv.org/content/early/2020/05/14/2020.05.12.077776.
- Mining for Strong Gravitational Lenses with Self-supervised Learning. The Astrophysical Journal, 932(2):107, 2022. ISSN 0004-637X. doi: 10.3847/1538-4357/ac6d63. URL http://arxiv.org/abs/2110.00023.
- Towards Galaxy Foundation Models with Hybrid Contrastive Learning. Machine Learning for Astrophysics Workshop at the Thirty Ninth International Conference on Machine Learning (ICML 2022), 6 2022b. doi: 10.48550/arxiv.2206.11927.