Galaxy Zoo: probabilistic morphology through Bayesian CNNs and active learning

Walmsley, Mike; Smith, Lewis; Lintott, Chris; Gal, Yarin; Bamford, Steven; Dickinson, Hugh; Fortson, Lucy; Kruk, Sandor; Masters, Karen; Scarlata, Claudia; Simmons, Brooke; Smethurst, Rebecca; Wright, Darryl

Abstract

We use Bayesian convolutional neural networks and a novel generative model of Galaxy Zoo volunteer responses to infer posteriors for the visual morphology of galaxies. Bayesian CNN can learn from galaxy images with uncertain labels and then, for previously unlabelled galaxies, predict the probability of each possible label. Our posteriors are well-calibrated (e.g. for predicting bars, we achieve coverage errors of 11.8 per cent within a vote fraction deviation of 0.2) and hence are reliable for practical use. Further, using our posteriors, we apply the active learning strategy BALD to request volunteer responses for the subset of galaxies which, if labelled, would be most informative for training our network. We show that training our Bayesian CNNs using active learning requires up to 35–60 per cent fewer labelled galaxies, depending on the morphological feature being classified. By combining human and machine intelligence, Galaxy zoo will be able to classify surveys of any conceivable scale on a time-scale of weeks, providing massive and detailed morphology catalogues to support research into galaxy evolution.

Citation

Walmsley, M., Smith, L., Lintott, C., Gal, Y., Bamford, S., Dickinson, H., …Wright, D. (2020). Galaxy Zoo: probabilistic morphology through Bayesian CNNs and active learning. Monthly Notices of the Royal Astronomical Society, 491(2), 1554-1574. https://doi.org/10.1093/mnras/stz2816

• Journal Article Type: Article
• Acceptance Date: Sep 27, 2019
• Online Publication Date: Oct 7, 2019
• Publication Date: Jan 11, 2020
• Deposit Date: Jan 16, 2020
• Publicly Available Date: Jan 16, 2020
• Journal: Monthly Notices of the Royal Astronomical Society
• Print ISSN: 0035-8711
• Electronic ISSN: 1365-2966
• Publisher: Oxford University Press
• Peer Reviewed: Peer Reviewed
• Volume: 491
• Issue: 2
• Pages: 1554-1574
• DOI: https://doi.org/10.1093/mnras/stz2816
• Keywords: Space and Planetary Science; Astronomy and Astrophysics
• Public URL: https://nottingham-repository.worktribe.com/output/3737384
• Publisher URL: https://academic.oup.com/mnras/article/491/2/1554/5583078
• Additional Information: This article has been accepted for publication in Monthly Notices of the Royal Astronomical Society (©: 2019 The authors) Published by Oxford University Press on behalf of the Royal Astronomical Society. All rights reserved.