arXiv subjects

Steven Bamford

Publications and source records attributed to Steven Bamford.

At least 19 recordsLinked to original sources

arXiv2025

Galaxy Zoo Evo: 1 million human-annotated images of galaxies

We introduce Galaxy Zoo Evo, a labeled dataset for building and evaluating foundation models on images of galaxies. GZ Evo includes 104M crowdsourced labels for 823k images from four telescopes. Each image is labeled with a series of fine-grained questions and answers (e.g. "featured galaxy, two spiral arms, tightly wound, merging with another galaxy"). These detailed labels are useful for pretraining or finetuning. We also include four smaller sets of labels (167k galaxies in total) for downstream tasks of specific interest to astronomers, including finding strong lenses and describing galaxies from the new space telescope Euclid. We hope GZ Evo will serve as a real-world benchmark for computer vision topics such as domain adaption (from terrestrial to astronomical, or between telescopes) or learning under uncertainty from crowdsourced labels. We also hope it will support a new generation of foundation models for astronomy; such models will be critical to future astronomers seeking to better understand our universe.

astro-ph.IM↗

arXiv2024

The Galaxy Zoo Catalogs for the Galaxy And Mass Assembly (GAMA) Survey

Galaxy Zoo is an online project to classify morphological features in extra-galactic imaging surveys with public voting. In this paper, we compare the classifications made for two different surveys, the Dark Energy Spectroscopic Instrument (DESI) imaging survey and a part of the Kilo-Degree Survey (KiDS), in the equatorial fields of the Galaxy And Mass Assembly (GAMA) survey. Our aim is to cross-validate and compare the classifications based on different imaging quality and depth. We find that generally the voting agrees globally but with substantial scatter i.e. substantial differences for individual galaxies. There is a notable higher voting fraction in favor of ``smooth'' galaxies in the DESI+\rev{\sc zoobot} classifications, most likely due to the difference between imaging depth. DESI imaging is shallower and slightly lower resolution than KiDS and the Galaxy Zoo images do not reveal details such as disk features \rev{and thus are missed in the {\sc zoobot} training sample}. \rev{We check against expert visual classifications and find good agreement with KiDS-based Galaxy Zoo voting.} We reproduce the results from Porter-Temple+ (2022), on the dependence of stellar mass, star-formation, and specific star-formation on the number of spiral arms. This shows that once corrected for redshift, the DESI Galaxy Zoo and KiDS Galaxy Zoo classifications agree well on population properties. The zoobot cross-validation increases confidence in its ability to compliment Galaxy Zoo classifications and its ability for transfer learning across surveys.

astro-ph.GA↗

arXiv2024

A Robust Study of High-Redshift Galaxies: Unsupervised Machine Learning for Characterising morphology with JWST up to z ~ 8

Galaxy morphologies provide valuable insights into their formation processes, tracing the spatial distribution of ongoing star formation and encoding signatures of dynamical interactions. While such information has been extensively investigated at low redshift, it is crucial to develop a robust system for characterising galaxy morphologies at earlier cosmic epochs. Relying solely on the nomenclature established for low-redshift galaxies risks introducing biases that hinder our understanding of this new regime. In this paper, we employ variational auto-encoders to perform feature extraction on galaxies at z $>$ 2 using JWST/NIRCam data. Our sample comprises 6869 galaxies at z $>$ 2, including 255 galaxies z $>$ 5, which have been detected in both the CANDELS/HST fields and CEERS/JWST, ensuring reliable measurements of redshift, mass, and star formation rates. To address potential biases, we eliminate galaxy orientation and background sources prior to encoding the galaxy features, thereby constructing a physically meaningful feature space. We identify 11 distinct morphological classes that exhibit clear separation in various structural parameters, such as CAS-$M_{20}$, Sérsic indices, specific star formation rates, and axis ratios. We observe a decline in the presence of spheroidal-type galaxies with increasing redshift, indicating a dominance of disk-like galaxies in the early universe. We demonstrate that conventional visual classification systems are inadequate for high-redshift morphology classification and advocate the need for a more detailed and refined classification scheme. Leveraging machine-extracted features, we propose a solution to this challenge and illustrate how our extracted clusters align with measured parameters, offering greater physical relevance compared to traditional methods.

astro-ph.GA↗

arXiv2023

Exploring the Morphologies of High Redshift Galaxies with Machine Learning

The morphology of a galaxy has been shown to encode the evolutionary history and correlates strongly with physical properties such as stellar mass, star formation rates and past merger events. While the majority of galaxies in the local universe can be classified on the Hubble sequence, little is known about the different types of galaxies we observe at high redshift. The irregular morphology of these galaxies makes visual classifications difficult, and with the future of astronomy consisting of many "Big Data" surveys we need an efficient, and unbiased classification system in place. In this work we explore the use of unsupervised machine learning techniques to preform feature extraction from galaxy images to separate high redshift galaxies into different morphological types based on the machine learning clusters. We expand on previous work by addressing observation biases such as the orientation, apparent size of the galaxies and noise before extracting features, thus reducing the number of clusters and forcing the network to learn meaningful features. We then compare the extracted clusters' physical properties, to investigate the separation between the groups.

astro-ph.GA↗

arXiv2022

Galaxy Zoo DECaLS: Detailed Visual Morphology Measurements from Volunteers and Deep Learning for 314,000 Galaxies

We present Galaxy Zoo DECaLS: detailed visual morphological classifications for Dark Energy Camera Legacy Survey images of galaxies within the SDSS DR8 footprint. Deeper DECaLS images (r=23.6 vs. r=22.2 from SDSS) reveal spiral arms, weak bars, and tidal features not previously visible in SDSS imaging. To best exploit the greater depth of DECaLS images, volunteers select from a new set of answers designed to improve our sensitivity to mergers and bars. Galaxy Zoo volunteers provide 7.5 million individual classifications over 314,000 galaxies. 140,000 galaxies receive at least 30 classifications, sufficient to accurately measure detailed morphology like bars, and the remainder receive approximately 5. All classifications are used to train an ensemble of Bayesian convolutional neural networks (a state-of-the-art deep learning method) to predict posteriors for the detailed morphology of all 314,000 galaxies. When measured against confident volunteer classifications, the networks are approximately 99% accurate on every question. Morphology is a fundamental feature of every galaxy; our human and machine classifications are an accurate and detailed resource for understanding how galaxies evolve.

astro-ph.GA↗

arXiv2021

A model-independent reconstruction of dark energy to very high redshift

We provide a model-independent reconstruction of dark energy from $z=0$ to $ \gtrsim 10^5$. We parameterise the model by a perfect fluid with a series of physically well-motivated bins in energy-density, such that the equation of state is always $-1 \le w \le 1$. Our method is capable of describing a range of theoretical models with smooth modifications to the expansion history. Combining the latest CMB, BAO, SN and local $H_0$ measurements, we obtain a large improvement of $Δχ^2=41.3$ over LCDM, at the expense of 33 additional parameters in the fit, with dark energy contributing significantly between $z \sim 10^4 - 10^5$, and intriguingly with a sound speed $c_s^2 \sim 1/3$. A significant part of the $Δχ^2$ improvement comes from \Planck\ + Atacama Cosmology Telescope (\textsc{Act}) data, alleviating tension between them within LCDM. We apply a correlation prior to penalise models with unnecessary degrees of freedom and find no preference for deviations from LCDM at late-times, but moderate Bayesian evidence of an early dark energy (EDE) component. Although the model has a large amount of freedom, it is unable to reduce $S_8 \equiv σ_8 (Ω_\mathrm{c} / 0.3)^{0.5}$ below that of \lcdm, to bring about full concordance with large-scale structure data.

astro-ph.CO↗

arXiv2021

Recovering the origins of the lenticular galaxy NGC 3115 using multi-band imaging

A detailed study of the morphology of lenticular galaxies is an important way to understand how this type of galaxy formed and evolves over time. Decomposing a galaxy into its components (disc, bulge, bar, ...) allows recovering the colour gradients present in each system, its star formation history, and its assembly history. We use GALFITM to perform a multi-wavelength structural decomposition of the closest lenticular galaxy, NGC 3115, resulting in the description of its stellar light into several main components: a bulge, a thin disc, a thick disc and also evidence of a bar. We report the finding of central bluer stellar populations in the bulge, as compared to the colour of the galaxy outskirts, indicating either the presence of an Active Galactic Nucleus (AGN) and/or recent star formation activity. From the spectral energy distribution results, we show that the galaxy has a low luminosity AGN component, but even excluding the effect of the nuclear activity, the bulge is still bluer than the outer-regions of the galaxy, revealing a recent episode of star formation. Based on all of the derived properties, we propose a scenario for the formation of NGC 3115 consisting of an initial gas-rich merger, followed by accretions and feedback that quench the galaxy, until a recent encounter with the companion KK084 that reignited the star formation in the bulge, provoked a core displacement in NGC 3115 and generated spiral-like features. This result is consistent with the two-phase formation scenario, proposed in previous studies of this galaxy.

astro-ph.GA↗

arXiv2020

Galaxy Zoo Builder: Four Component Photometric decomposition of Spiral Galaxies Guided by Citizen Science

Multi-component modelling of galaxies is a valuable tool in the effort to quantitatively understand galaxy evolution, yet the use of the technique is plagued by issues of convergence, model selection and parameter degeneracies. These issues limit its application over large samples to the simplest models, with complex models being applied only to very small samples. We attempt to resolve this dilemma of "quantity or quality" by developing a novel framework, built inside the Zooniverse citizen science platform, to enable the crowdsourcing of model creation for Sloan Digitial Sky Survey galaxies. We have applied the method, including a final algorithmic optimisation step, on a test sample of 198 galaxies, and examine the robustness of this new method. We also compare it to automated fitting pipelines, demonstrating that it is possible to consistently recover accurate models that either show good agreement with, or improve on, prior work. We conclude that citizen science is a promising technique for modelling images of complex galaxies, and release our catalogue of models.

astro-ph.GA↗

arXiv2019

Galaxy Zoo: Probabilistic Morphology through Bayesian CNNs and Active Learning

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% 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% 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.

astro-ph.GA↗

arXiv2018

Galaxy Zoo: Morphological classification of galaxy images from the Illustris simulation

Modern cosmological simulations model the universe with increasing sophistication and at higher spatial and temporal resolutions. These enhancements permit detailed comparisons between the simulation outputs and real observational data. Recent projects such as Illustris are capable of producing simulated images that are comparable to those obtained from local surveys. This paper tests how well Illustris achieves this goal across a diverse population of galaxies using visual morphologies derived from Galaxy Zoo citizen scientists. Morphological classifications provided by volunteers for simulated galaxies are compared with similar data for a compatible sample of images drawn from the SDSS Legacy Survey. This paper investigates how simple morphological characterization by human volunteers asked to distinguish smooth from featured systems differs between simulated and real galaxy images. Differences are identified, which are likely due to the limited resolution of the simulation, but which could be revealing real differences in the dynamical evolution of populations of galaxies in the real and model universes. Specifically, for stellar masses $M_{\star}\lesssim10^{11}M_{\odot}$, a larger proportion of Illustris galaxies that exhibit disk-like morphology or visible substructure, relative to their SDSS counterparts. Toward higher masses, simulated and observed galaxies converge and exhibit similar morphology distributions. The stellar mass threshold indicated by this divergent behavior confirms recent works using parametric measures of morphology from Illustris simulated images. When $M_{\star}\gtrsim10^{11}M_{\odot}$, the Illustris dataset contains fewer galaxies that classifiers regard as unambiguously featured. These results suggest that comparison between the detailed properties of observed and simulated galaxies, even when limited to reasonably massive systems, may be misleading.

astro-ph.GA↗

arXiv2016

SDSS-IV MaNGA: Bulge-Disc Decomposition of IFU Datacubes (BUDDI)

With the availability of large integral-field unit (IFU) spectral surveys of nearby galaxies, there is now the potential to extract spectral information from across the bulges and discs of galaxies in a systematic way. This information can address questions such as how these components built up with time, how galaxies evolve and whether their evolution depends on other properties of the galaxy such as its mass or environment. We present BUDDI, a new approach to fit the two-dimensional light profiles of galaxies as a function of wavelength to extract the spectral properties of these galaxies' discs and bulges. The fitting is carried out using GalfitM, a modified form of Galfit which can fit multi-waveband images simultaneously. The benefit of this technique over traditional multi-waveband fits is that the stellar populations of each component can be constrained using knowledge over the whole image and spectrum available. The decomposition has been developed using commissioning data from the SDSS-IV Mapping Nearby Galaxies at APO (MaNGA) survey with redshifts z<0.14 and coverage of at least 1.5 effective radii for a spatial resolution of 2.5 arcsec FWHM and field of view of >22 arcsec, but can be applied to any IFU data of a nearby galaxy with similar or better spatial resolution and coverage. We present an overview of the fitting process, the results from our tests, and we finish with example stellar population analyses of early-type galaxies from the MaNGA survey to give an indication of the scientific potential of applying bulge-disc decomposition to IFU data.

astro-ph.GA↗

arXiv2016

Galaxy Zoo: Mergers - Dynamical Models of Interacting Galaxies

The dynamical history of most merging galaxies is not well understood. Correlations between galaxy interaction and star formation have been found in previous studies, but require the context of the physical history of merging systems for full insight into the processes that lead to enhanced star formation. We present the results of simulations that reconstruct the orbit trajectories and disturbed morphologies of pairs of interacting galaxies. With the use of a restricted three-body simulation code and the help of Citizen Scientists, we sample 10^5 points in parameter space for each system. We demonstrate a successful recreation of the morphologies of 62 pairs of interacting galaxies through the review of more than 3 million simulations. We examine the level of convergence and uniqueness of the dynamical properties of each system. These simulations represent the largest collection of models of interacting galaxies to date, providing a valuable resource for the investigation of mergers. This paper presents the simulation parameters generated by the project. They are now publicly available in electronic format at http://data.galaxyzoo.org/mergers.html. Though our best-fit model parameters are not an exact match to previously published models, our method for determining uncertainty measurements will aid future comparisons between models. The dynamical clocks from our models agree with previous results of the time since the onset of star formation from star burst models in interacting systems and suggests that tidally induced star formation is triggered very soon after closest approach.

astro-ph.GA↗

arXiv2016

The Moon Zoo citizen science project: Preliminary results for the Apollo 17 landing site

Moon Zoo is a citizen science project that utilises internet crowd-sourcing techniques. Moon Zoo users are asked to review high spatial resolution images from the Lunar Reconnaissance Orbiter Camera (LROC), onboard NASAs LRO spacecraft, and perform characterisation such as measuring impact crater sizes and identify morphological features of interest. The tasks are designed to address issues in lunar science and to aid future exploration of the Moon. We have tested various methodologies and parameters therein to interrogate and reduce the Moon Zoo crater location and size dataset against a validated expert survey. We chose the Apollo 17 region as a test area since it offers a broad range of cratered terrains, including secondary-rich areas, older maria, and uplands. The assessment involved parallel testing in three key areas: (1) filtering of data to remove problematic mark-ups; (2) clustering methods of multiple notations per crater; and (3) derivation of alternative crater degradation indices, based on the statistical variability of multiple notations and the smoothness of local image structures. We compared different combinations of methods and parameters and assessed correlations between resulting crater summaries and the expert census. We derived the optimal data reduction steps and settings of the existing Moon Zoo crater data to agree with the expert census. Further, the regolith depth and crater degradation states derived from the data are also found to be in broad agreement with other estimates for the Apollo 17 region. Our study supports the validity of this citizen science project but also recommends improvements in key elements of the data acquisition planning and production.

astro-ph.EP↗

arXiv2015

Galaxy And Mass Assembly (GAMA): Trends in galaxy colours, morphology, and stellar populations with large scale structure, group, and pair environments

We explore trends in galaxy properties with Mpc-scale structures using catalogues of environment and large scale structure from the Galaxy And Mass Assembly (GAMA) survey. Existing GAMA catalogues of large scale structure, group and pair membership allow us to construct galaxy stellar mass functions for different environmental types. To avoid simply extracting the known underlying correlations between galaxy properties and stellar mass, we create a mass matched sample of galaxies with stellar masses between $9.5 \leq \log{M_*/h^{-2} M_{\odot}} \leq 11$ for each environmental population. Using these samples, we show that mass normalised galaxies in different large scale environments have similar energy outputs, $u-r$ colours, luminosities, and morphologies. Extending our analysis to group and pair environments, we show galaxies that are not in groups or pairs exhibit similar characteristics to each other regardless of broader environment. For our mass controlled sample, we fail to see a strong dependence of Sérsic index or galaxy luminosity on halo mass, but do find that it correlates very strongly with colour. Repeating our analysis for galaxies that have not been mass controlled introduces and amplifies trends in the properties of galaxies in pairs, groups, and large scale structure, indicating that stellar mass is the most important predictor of the galaxy properties we examine, as opposed to environmental classifications.

astro-ph.GA↗

arXiv2014

Galaxy And Mass Assembly (GAMA): mass-size relations of z$<$0.1 galaxies subdivided by Sérsic index, colour and morphology

We use data from the Galaxy And Mass Assembly (GAMA) survey in the redshift range 0.01$<$z$<$0.1 (8399 galaxies in $g$ to $K_s$ bands) to derive the stellar mass $-$ half-light radius relations for various divisions of 'early' and 'late'-type samples. We find the choice of division between early and late (i.e., colour, shape, morphology) is not particularly critical, however, the adopted mass limits and sample selections (i.e., the careful rejection of outliers and use of robust fitting methods) are important. In particular we note that for samples extending to low stellar mass limits ($<10^{10}\mathcal{M_{\odot}}$) the Sérsic index bimodality, evident for high mass systems, becomes less distinct and no-longer acts as a reliable separator of early- and late-type systems. The final set of stellar mass $-$ half-light radius relations are reported for a variety of galaxy population subsets in 10 bands ($ugrizZYJHK_s$) and are intended to provide a comprehensive low-z benchmark for the many ongoing high-z studies. Exploring the variation of the stellar mass $-$ half-light radius relations with wavelength we confirm earlier findings that galaxies appear more compact at longer wavelengths albeit at a smaller level than previously noted: at $10^{10}\mathcal{M_{\odot}}$ both spiral systems and ellipticals show a decrease in size of 13% from $g$ to $K_s$ (which is near linear in log wavelength). Finally we note that the sizes used in this work are derived from 2D Sérsic light profile fitting (using GALFIT3), i.e., elliptical semi-major half light radii, improving on earlier low-z benchmarks based on circular apertures.

astro-ph.GA↗

arXiv2014

Galaxy Zoo: CANDELS Barred Disks and Bar Fractions

The formation of bars in disk galaxies is a tracer of the dynamical maturity of the population. Previous studies have found that the incidence of bars in disks decreases from the local Universe to z ~ 1, and by z > 1 simulations predict that bar features in dynamically mature disks should be extremely rare. Here we report the discovery of strong barred structures in massive disk galaxies at z ~ 1.5 in deep rest-frame optical images from CANDELS. From within a sample of 876 disk galaxies identified by visual classification in Galaxy Zoo, we identify 123 barred galaxies. Selecting a sub-sample within the same region of the evolving galaxy luminosity function (brighter than L*), we find that the bar fraction across the redshift range 0.5< z < 2 (f_bar = 10.7 +6.3 -3.5% after correcting for incompleteness) does not significantly evolve. We discuss the implications of this discovery in the context of existing simulations and our current understanding of the way disk galaxies have evolved over the last 11 billion years.

astro-ph.GA↗

arXiv2013

Morphology in the Era of Large Surveys

The study of galaxies has changed dramatically over the past few decades with the advent of large-scale astronomical surveys. These large collaborative efforts have made available high-quality imaging and spectroscopy of hundreds of thousands of systems, providing a body of observations which has significantly enhanced our understanding not only of cosmology and large-scale structure in the universe but also of the astrophysics of galaxy formation and evolution. Throughout these changes, one thing that has remained constant is the role of galaxy morphology as a clue to understanding galaxies. But obtaining morphologies for large numbers of galaxies is challenging; this topic, "Morphology in the era of large surveys", was the subject of a recent discussion meeting at the Royal Astronomical Society, and this "Astronomy and Geophysics" article is a report on that meeting.

astro-ph.CO↗

arXiv2012

The effect of the environment on the gas kinematics and the structure of distant galaxies

With the aim of distinguishing between possible physical mechanisms acting on galaxies when they fall into clusters, we study the properties of the gas and the stars in a sample of 422 emission-line galaxies from EDisCS in different environments up to z~1. We identify galaxies with kinematical disturbances (from emission-lines in their 2D spectra) and find that they are more frequent in clusters than in the field. The fraction of kinematically-disturbed galaxies increases with cluster velocity dispersion and decreases with distance from the cluster centre, but remains constant with projected galaxy density. We also studied morphological disturbances in the stellar light from HST/F814W images, finding that the fraction of morphologically disturbed galaxies is similar in all environments. Moreover, there is little correlation between the presence of kinematically-disturbed gas and morphological distortions. We also study the dependence of the Tully-Fisher relation, star formation, and extent of the emission on environment, and conclude that the gas disks in cluster galaxies have been truncated, and therefore their star formation is more concentrated than in low-density environments. If spirals transform into S0s, our findings imply that the physical mechanism transforming cluster galaxies efficiently disturbs the star-forming gas and reduces their specific star formation rate. Moreover, this star-forming gas is either removed more efficiently from the outskirts of the galaxies or is driven towards the centre (or both), helping to build the bulges of S0s. These results, in addition to the finding that the transformation mechanism does not seem to induce strong morphological disturbances on the galaxies, suggest that the physical processes involved are related to the intracluster medium, with galaxy-galaxy interactions playing only a limited role in clusters.

astro-ph.CO↗