Galaxy Zoo: probabilistic morphology through Bayesian CNNs and active learning

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 10.6% within 5 responses and 2.9% within 10 responses) 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% 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 timescale of weeks, providing massive and detailed morphology catalogues to support research into galaxy evolution.

Machine-Learning Methods for Computational Science and Engineering

“…While ML methods such as ANN have been used for almost 30 years [226], more recent works focus on CNNs, due to their ability to process and analyze images in a relatively computationally efficient way. CNNs have been used to understand the morphology of galaxies [227][228][229], predict photometric redshifts [230,231], detect galaxy clusters [232], identify gravitational lenses [233][234][235][236] and reconstruction of images [237] Video classification is yet another field that keeps improving along with advances in ML. Karpathy et al [238] have used CNNs to classify sports-related videos found on YouTube into their corresponding sports.…”

A self-supervised, physics-aware, Bayesian neural network architecture for modelling galaxy emission-line kinematics

“…Gal (2016) first proposed the idea of approximating distributions over parameters learned in neural networks in this way and has since been used in astronomy (e.g. for the probabilistic labelling of galaxy morphologies, Walmsley et al 2019).…”

“…For a comprehensive derivation of Equations 6 and 7, as well as the implications for using an arbitrary dropout probability, we refer the reader to Walmsley et al (2019). Examples of the posterior distributions, p(k|w,D), over learned parameters using MC dropout for a randomly selected synthesised galaxy are described further in §3.1.1.…”

fink, a new generation of broker for the LSST community

“…Identifying such objects is the goal of active learning algorithms which have proven to be successful in a series of astronomical applications (Solorio et al 2005;Richards et al 2012;Ishida et al 2019b). Furthermore, Bayesian neural networks (BNNs) have been shown to provide meaningful classification uncertainties that are linked to the representativeness of training sets and can be used as information for the AL loop (Möller & de Boissière 2019;Walmsley et al 2020).…”