Dwarfs from the Dark (Energy Survey): a machine learning approach to classify dwarf galaxies from multi-band images

Müller, Oliver; Schnider, Eva

doi:10.48550/arxiv.2102.12776

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Over the past decade, machine learning and in particular CNNs have become more prominent following from advances in GPUs, the creation of large accessible datasets (Deng et al 2009), and the success of the deep learning CNN AlexNet (Krizhevsky et al 2012). As it has become popular in the mainstream, machine learning has been increasingly used in astronomy, in particular to classification problems, such as: using CNNs and multi-band images to classify dwarf galaxies (Müller & Schnider 2021), using CNNs to assign Fanaroff-Riley classifications to radio galaxies (Scaife & Porter 2021), using an autoencoder for morphological classification of galaxies (Spindler et al 2021), using a U-Net to perform source detection, segmentation, and classification (Hausen & Robertson 2020), using RNN to correct classifications in maps (Maggiori et al 2017), and classifying galaxies using T-SNE (Zhang et al 2020).…”

Section: Introductionmentioning

confidence: 99%

Detecting gravitational lenses using machine learning: exploring interpretability and sensitivity to rare lensing configurations

Wilde,

Serjeant,

Bromley

et al. 2022

Preprint

View full text Add to dashboard Cite

Forthcoming large imaging surveys such as Euclid and the Vera Rubin Observatory Legacy Survey of Space and Time are expected to find more than 10 5 strong gravitational lens systems, including many rare and exotic populations such as compound lenses, but these 10 5 systems will be interspersed among much larger catalogues of ∼ 10 9 galaxies. This volume of data is too much for visual inspection by volunteers alone to be feasible and gravitational lenses will only appear in a small fraction of these data which could cause a large amount of false positives. Machine learning is the obvious alternative but the algorithms' internal workings are not obviously interpretable, so their selection functions are opaque and it is not clear whether they would select against important rare populations. We design, build, and train several Convolutional Neural Networks (CNNs) to identify strong gravitational lenses using VIS, Y, J, and H bands of simulated data, with F1 scores between 0.83 and 0.91 on 100,000 test set images. We demonstrate for the first time that such CNNs do not select against compound lenses, obtaining recall scores as high as 76% for compound arcs and 52% for double rings. We verify this performance using Hubble Space Telescope (HST) and Hyper Suprime-Cam (HSC) data of all known compound lens systems. Finally, we explore for the first time the interpretability of these CNNs using Deep Dream, Guided Grad-CAM, and by exploring the kernels of the convolutional layers, to illuminate why CNNs succeed in compound lens selection. 1

show abstract