AstroVaDEr: astronomical variational deep embedder for unsupervised morphological classification of galaxies and synthetic image generation

Abstract

We present AstroVaDEr, a variational autoencoder designed to perform unsupervised clustering and synthetic image generation using astronomical imaging catalogues. The model is a convolutional neural network that learns to embed images into a low dimensional latent space, and simultaneously optimises a Gaussian Mixture Model (GMM) on the embedded vectors to cluster the training data. By utilising variational inference, we are able to use the learned GMM as a statistical prior on the latent space to facilitate random sampling and generation of synthetic images. We demonstrate AstroVaDEr’s capabilities by training it on gray-scaled gri images from the Sloan Digital Sky Survey, using a sample of galaxies that are classified by Galaxy Zoo 2. An unsupervised clustering model is found which separates galaxies based on learned morphological features such as axis ratio, surface brightness profile, orientation and the presence of companions. We use the learned mixture model to generate synthetic images of galaxies based on the morphological profiles of the Gaussian components. AstroVaDEr succeeds in producing a morphological classification scheme from unlabelled data, but unexpectedly places high importance on the presence of companion objects—demonstrating the importance of human interpretation. The network is scalable and flexible, allowing for larger datasets to be classified, or different kinds of imaging data. We also demonstrate the generative properties of the model, which allow for realistic synthetic images of galaxies to be sampled from the learned classification scheme. These can be used to create synthetic image catalogs or to perform image processing tasks such as deblending.

Self-supervised Representation Learning for Astronomical Images

“…GZ2 (Willett et al 2013), the successor to GZ, is focused on more fine-grained features and in total achieved morphological classifications of 304,122 SDSS galaxies. Shown most prominently by the winners of the "Galaxy Challenge" (Dieleman et al 2015) and numerous subsequent works since (Domínguez Sánchez et al 2018, 2019Khan et al 2019;Spindler et al 2020;Vega-Ferrero et al 2020;Walmsley et al 2020), CNNs excel at this task.…”

“…Unsupervised ML methods aim to learn semantically meaningful representations of the data without relying on any labels (see, e.g., Alloghani et al 2020). Many such methods have already been applied to studies of galaxy morphology (Hocking et al 2018;Cheng et al 2020a;Martin et al 2020;Spindler et al 2020), identification of strong lenses (Cheng et al 2020b), and anomaly detection (Xiong et al 2018;Margalef-Bentabol et al 2020). Unfortunately, across most computer vision applications, the utility of unsupervised representations for downstream tasks has historically lagged behind that of the representations coming from supervised training (Caron et al 2018).…”

Synthetic Data Generation to Speed-Up the Object Recognition Pipeline

“…Furthermore, synthetic scenario creation is becoming increasingly relevant in a variety of fields [29][30][31][32][33][34][35], ranging from an object and figure identification and categorisation to automated recognition and tracing [36][37][38][39]. The adaptability and reusability of this method is exemplified by the ability to train networks to detect specific items or people in very challenging circumstances.…”

A Survey on Variational Autoencoders from a Green AI Perspective

“…Variational Autoencoders are becoming increasingly popular inside the scientific community [53,60,61], both due to their strong probabilistic foundation, that will be recalled in "Theoretical Background", and the precious insight on the latent representation of data. However, in spite of the remarkable achievements, the behaviour of Variational Autoencoders is still far from satisfactory; there is a number of well-known theoretical and practical challenges that still hinder this generative paradigm (see "The Vanilla VAE and Its Problems"), and whose solution drove the recent research on this topic.…”