arXiv subjects

Ashley Spindler

Publications and source records attributed to Ashley Spindler.

7 recordsLinked to original sources

arXiv2026

Deep Learning Segmentation of Spiral Arms and Bars for 600,000 Galaxies in DESI

We present a catalogue of segmentation maps identifying the extent of spiral arms and bars of 639,636 galaxies in the DESI Legacy Survey. To produce these maps, we have trained a deep U-net-style neural network using the pixel masks from the Galaxy Zoo: 3D citizen science project. The resulting data products are "soft" segmentation maps, which show the confidence of the model that a pixel lies within the spiral arms or bars of a galaxy. In this paper we detail the sample selection from DESI-LS using the machine classifications from Galaxy Zoo: DESI, the architecture of the U-net model--dubbed ZooBot:3D. We demonstrate the ability of the model to identify spiral arms and bars in a wide range of face-on disks, and identify an emergent ability to identify rings--despite only a small number of ring-type galaxies being present in the training data. Finally, we discuss the practical application of these data products to photometric imaging and IFU spectroscopy. The ZooBot:3D dataset is available for use publicly, and contains the full catalogue presented in this paper, along with cross-matched subsamples for the MaNGA and SAMI IFU surveys.

astro-ph.GA↗

arXiv2026

The effects of image augmentations when training machine learning models in astronomy

We measure the influence of image augmentations and training dataset size when training a deep neural network to classify galaxy morphology. Data augmentation is an integral step when training machine learning models and often astronomers add augmentations assuming they will always improve the performance of their models. We train multiple versions of the same pre-existing Zoobot model using different image augmentations and different dataset sizes from 230,000 galaxy images from Galaxy Zoo DECaLS to determine whether this assumption is necessarily true. We find that generally, the addition of image augmentations does improve a deep neural network's performance, however, this improvement is significantly diminished as the training dataset size increases. The choice of specific augmentations (provided they are sensible) does not seem to be as important as simply having augmentations as different augmentations result in similar increases in performances. We find that for a model of a given size, there exists a saturation point (when the model's capacity has been filled with data) that cannot be surpassed with data augmentations. We find that more complex augmentations result in longer training times and might not lead to improved performance. If augmentations are added to the training process (which is recommended), simpler augmentations might be sufficient, depending on the size of the dataset and model. We therefore encourage astronomers to carefully consider their use of image augmentations in an effort to reduce wasted time and computational resources.

astro-ph.IM↗

arXiv2023

Deep Learning Segmentation of Spiral Arms and Bars

We present the first deep learning model for segmenting galactic spiral arms and bars. In a blinded assessment by expert astronomers, our predicted spiral arm masks are preferred over both current automated methods (99% of evaluations) and our original volunteer labels (79% of evaluations). Experts rated our spiral arm masks as `mostly good' to `perfect' in 89% of evaluations. Bar lengths trivially derived from our predicted bar masks are in excellent agreement with a dedicated crowdsourcing project. The pixelwise precision of our masks, previously impossible at scale, will underpin new research into how spiral arms and bars evolve.

astro-ph.GA↗

arXiv2020

AstroVaDEr: Astronomical Variational Deep Embedder for Unsupervised Morphological Classification of Galaxies and Synthetic Image Generation

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 \textit{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.

astro-ph.IM↗

arXiv2018

SDSS IV MaNGA - sSFR profiles and the slow quenching of discs in green valley galaxies

We study radial profiles in H$α$ equivalent width and specific star formation rate (sSFR) derived from spatially-resolved SDSS-IV MaNGA spectroscopy to gain insight on the physical mechanisms that suppress star formation and determine a galaxy's location in the SFR-$\rm M_\star$ diagram. Even within the star-forming `main sequence', the measured sSFR decreases with stellar mass, both in an integrated and spatially-resolved sense. Flat sSFR radial profiles are observed for $\rm log(M_\star/ M_\odot) < 10.5$, while star-forming galaxies of higher mass show a significant decrease in sSFR in the central regions, a likely consequence of both larger bulges and an inside-out growth history. Our primary focus is the green valley, constituted by galaxies lying below the star formation main sequence, but not fully passive. In the green valley we find sSFR profiles that are suppressed with respect to star-forming galaxies of the same mass at all galactocentric distances out to 2 effective radii. The responsible quenching mechanism therefore appears to affect the entire galaxy, not simply an expanding central region. The majority of green valley galaxies of $\rm log(M_\star/ M_\odot) > 10.0$ are classified spectroscopically as central low-ionisation emission-line regions (cLIERs). Despite displaying a higher central stellar mass concentration, the sSFR suppression observed in cLIER galaxies is not simply due to the larger mass of the bulge. Drawing a comparison sample of star forming galaxies with the same $\rm M_\star$ and $\rm Σ_{1~kpc}$ (the mass surface density within 1 kpc), we show that a high $\rm Σ_{1~kpc}$ is not a sufficient condition for determining central quiescence.

astro-ph.GA↗

arXiv2018

SDSS-IV MaNGA: The Spatial Distribution of Star Formation and its Dependence on Mass, Structure and Environment

We study the spatially resolved star formation of 1494 galaxies in the SDSSIV-MaNGA Survey. SFRs are calculated using a two-step process, using $H_α$ in star forming regions and $D_n4000$ in regions identified as AGN/LI(N)ER or lineless. The roles of secular and environmental quenching processes are investigated by studying the dependence of the radial profiles of specific star formation rate on stellar mass, galaxy structure and environment. We report on the existence of `Centrally Suppressed' galaxies, which have suppressed SSFR in their cores compared to their disks. The profiles of centrally suppressed and unsuppressed galaxies are distibuted in a bimodal way. Galaxies with high stellar mass and core velocity dispersion are found to be much more likely to be centrally suppressed than low mass galaxies, and we show that this is related to morphology and the presence of AGN/LI(N)ER like emission. Centrally suppressed galaxies also display lower star formation at all radii compared to unsuppressed galaxies. The profiles of central and satellite galaxies are also compared, and we find that satellite galaxies experience lower specific star formation rates at all radii than central galaxies. This uniform suppression could be a signal of the stripping of hot halo gas in the process known as strangulation. We find that satellites are not more likely to be suppressed in their cores than centrals, indicating that the core suppression is an entirely internal process. We find no correlation between the local environment density and the profiles of star formation rate surface density.

astro-ph.GA↗

arXiv2017

The Differing Relationships Between Size, Mass, Metallicity and Core Velocity Dispersion of Central and Satellite Galaxies

We study the role of environment in the evolution of central and satellite galaxies with the Sloan Digital Sky Survey. We begin by studying the size-mass relation, replicating previous studies, which showed no difference between the sizes of centrals and satellites at fixed stellar mass, before turning our attention to the size-core velocity dispersion ($σ_0$) and mass-$σ_0$ relations. By comparing the median size and mass of the galaxies at fixed velocity dispersion we find that the central galaxies are consistently larger and more massive than their satellite counterparts in the quiescent population. In the star forming population we find there is no difference in size and only a small difference in mass. To analyse why these difference may be present we investigate the radial mass profiles and stellar metallicity of the galaxies. We find that in the cores of the galaxies there is no difference in mass surface density between centrals and satellites, but there is a large difference at larger radii. We also find almost no difference between the stellar metallicity of centrals and satellites when they are separated into star forming and quiescent groups. Under the assumption that $σ_0$ is invariant to environmental processes, our results imply that central galaxies are likely being increased in mass and size by processes such as minor mergers, particularly at high $σ_0$, while satellites are being slightly reduced in mass and size by tidal stripping and harassment, particularly at low $σ_0$, all of which predominantly affect the outer regions of the galaxies.

astro-ph.GA↗