Morphological classification of Radio Galaxies using Semi-Supervised Group Equivariant CNNs

Abstract

Out of the estimated few trillion galaxies, only around a million have been detected through radio frequencies, and only a tiny fraction, approximately a thousand, have been manually classified. We have addressed this disparity between labeled and unlabeled images of radio galaxies by employing a semi-supervised learning approach to classify them into the known Fanaroff-Riley Type I (FRI) and Type II (FRII) categories. A Group Equivariant Convolutional Neural Network (G-CNN) was used as an encoder of the state-of-the-art self-supervised methods SimCLR (A Simple Framework for Contrastive Learning of Visual Representations) and BYOL (Bootstrap Your Own Latent). The G-CNN preserves the equivariance for the Euclidean Group E(2), enabling it to effectively learn the representation of globally oriented feature maps. After representation learning, we trained a fully-connected classifier and fine-tuned the trained encoder with labeled data. Our findings demonstrate that our semi-supervised approach outperforms existing state-of-the-art methods across several metrics, including cluster quality, convergence rate, accuracy, precision, recall, and the F1-score. Moreover, statistical significance testing via a t-test revealed that our method surpasses the performance of a fully supervised G-CNN. This study emphasizes the importance of semi-supervised learning in radio galaxy classification, where labeled data are still scarce, but the prospects for discovery are immense.

CCDS Authors

M Ashraful Amin, PhD

Founder DirectorDirector, Artificial Intelligence & Machine Learing WingDirector, Human-Computer Interaction WingDirector, Data Science Wing

References

1. 1.B. L. Fanaroff, J. M. Riley. (1974). The Morphology of Extragalactic Radio Sources of High and Low Luminosity. Monthly Notices of the Royal Astronomical Society, 167(1), 31P–36P[10.1093/mnras/167.1.31p]
2. 2.A. K. Aniyan, K. Thorat. (2017). Classifying Radio Galaxies with the Convolutional Neural Network. The Astrophysical Journal Supplement Series, 230(2), 20[10.3847/1538-4365/aa7333]
3. 3.Julie Banfield, O. Ivy Wong, Kyle Willett, R. P. Norris, L. Rudnick, Stanislav S. Shabala, Brooke Simmons, C Snyder, Avery Garon, N. Seymour, E. Middelberg, H. Andernach, Chris Lintott, Kristin A. Jacob, A. D. Kapińska, Minnie Mao, Karen L. Masters, M. J. Jarvis, Kevin Schawinski, Edward Paget, R. J. Simpson, H.-R. Klöckner, S. P. Bamford, T. Burchell, Kate Chow, Garret Cotter, L. Fortson, Ian Heywood, T. W. Jones, Sugata Kaviraj, Á. R. López-Sánchez, W. Peter Maksym, Kai Polsterer, K. Borden, Robert Hollow, L. Whyte. (2015). Radio Galaxy Zoo: host galaxies and radio morphologies derived from visual inspection. Monthly Notices of the Royal Astronomical Society, 453(3), 2327–2341[10.1093/mnras/stv1688]
4. 4.Chen Wu, O. Ivy Wong, L. Rudnick, Stanislav S. Shabala, Matthew J. Alger, Julie Banfield, Cheng Soon Ong, S. V. White, Avery Garon, R. P. Norris, H. Andernach, Jean Tate, Vesna Lukic, Hongming Tang, Kevin Schawinski, Foivos I. Diakogiannis. (2018). Radio Galaxy Zoo:<scp>Claran</scp>– a deep learning classifier for radio morphologies. Monthly Notices of the Royal Astronomical Society, 482(1), 1211–1230[10.1093/mnras/sty2646]
5. 5.Halime Miraghaei, P. N. Best. (2017). The nuclear properties and extended morphologies of powerful radio galaxies: the roles of host galaxy and environment. Monthly Notices of the Royal Astronomical Society, stx007[10.1093/mnras/stx007]
6. 6.D. D. Proctor. (2011). MORPHOLOGICAL ANNOTATIONS FOR GROUPS IN THE FIRST DATABASE. The Astrophysical Journal Supplement Series, 194(2), 31[10.1088/0067-0049/194/2/31]
7. 7.Zhixian Ma, Haiguang Xu, Jie Zhu, Dan Hu, Weitian Li, Chenxi Shan, Zhenghao Zhu, Liyi Gu, Jinjin Li, Chengze Liu, Xiangping Wu. (2019). A Machine Learning Based Morphological Classification of 14,245 Radio AGNs Selected from the Best–Heckman Sample. The Astrophysical Journal Supplement Series, 240(2), 34[10.3847/1538-4365/aaf9a2]
8. 8.Timothy J. Galvin, Minh Huynh, R. P. Norris, X. Rosalind Wang, Ellen Hopkins, Kai Lars Polsterer, Nicholas Ralph, Andrew O’Brien, G. Heald. (2020). Cataloguing the radio-sky with unsupervised machine learning: a new approach for the SKA era. Monthly Notices of the Royal Astronomical Society, 497(3), 2730–2758[10.1093/mnras/staa1890]
9. 9.Wathela Alhassan, A. R. Taylor, M. Vaccari. (2018). The FIRST Classifier: compact and extended radio galaxy classification using deep Convolutional Neural Networks. Monthly Notices of the Royal Astronomical Society, 480(2), 2085–2093[10.1093/mnras/sty2038]
10. 10.Anna M. M. Scaife, Fiona Porter. (2021). Fanaroff–Riley classification of radio galaxies using group-equivariant convolutional neural networks. Monthly Notices of the Royal Astronomical Society, 503(2), 2369–2379[10.1093/mnras/stab530]
11. 11.Inigo Val Slijepcevic, Anna M. M. Scaife, Mike Walmsley, Micah Bowles, O. Ivy Wong, Stanislav S. Shabala, Hongming Tang. (2022). Radio Galaxy Zoo: Using semi-supervised learning to leverage large unlabelled data-sets for radio galaxy classification under data-set shift. arXiv (Cornell University)[10.1093/mnras/stac1135]
12. 12.Tapan K. Sasmal, Soumen Bera, Sabyasachi Pal, Soumen Mondal. (2022). A New Catalog of Head–Tail Radio Galaxies from the VLA FIRST Survey. The Astrophysical Journal Supplement Series, 259(2), 31[10.3847/1538-4365/ac4473]
13. 13.Zhixian Ma, Jie Zhu, Yongkai Zhu, Haiguang Xu. (2019). Classification of Radio Galaxy Images with Semi-supervised Learning. Communications in computer and information science, 191–200[10.1007/978-981-32-9563-6_20]
14. 14.Jean-Bastien Grill, Florian Strub, Florent Altché, Corentin Tallec, Pierre H. Richemond, Elena Buchatskaya, Carl Doersch, Bernardo Ávila Pires, Zhaohan Daniel Guo, Mohammad Gheshlaghi Azar, Bilal Piot, Koray Kavukcuoglu, Rémi Munos, Michal Vaľko. (2020). Bootstrap your own latent: A new approach to self-supervised Learning. arXiv (Cornell University)[10.48550/arxiv.2006.07733]
15. 15.Fiona Porter. (2020). MiraBest Batched Dataset. Zenodo (CERN European Organization for Nuclear Research)[10.5281/zenodo.4288837]