Radio Galaxy Zoo: Unsupervised Clustering of Convolutionally Auto-encoded Radio-astronomical Images

Ralph, Nicholas O; Norris, Ray; Fang, Gengfa; Park, Laurence A. F.; Galvin, T. J.; Alger, Matthew J.; Andernach, H.; Lintott, Chris; Rudnick, Lawrence; Shabala, Stanislav S.; Wong, O. I.

doi:10.1088/1538-3873/ab213d

Cited by 53 publications

(33 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such tools can also be connected to automatic ways of exploring the morphological classification of radio sources, which are now starting to be developed using automatic techniques (see e.g. Mingo et al 2019;Ralph et al 2019;Galvin et al 2019;Mostert et al 2020).…”

Section: Conclusion and Future Outlookmentioning

confidence: 99%

The best of both worlds: Combining LOFAR and Apertif to derive resolved radio spectral index images

Morganti

Oosterloo

Brienza

et al. 2021

A&A

View full text Add to dashboard Cite

Super massive black holes at the centres of galaxies can cycle through periods of activity and quiescence. Characterising the duty cycle of active galactic nuclei (AGN) is crucial for understanding the impact of the energy they release on the host galaxy. For radio AGN, this can be done by identifying dying (remnant) and restarted radio galaxies from their radio spectral properties. Using the combination of the images at 1400 MHz produced by Apertif, the new phased-array feed receiver installed on the Westerbork Synthesis Radio Telescope, and images at 150 MHz provided by LOFAR, we have derived resolved spectral index images (at a resolution of ~15 arcsec) for all the sources within an approximately 6 deg2 area of the Lockman Hole region. In this way, we were able to select 15 extended radio sources with emission (partly or entirely) characterised by extremely steep spectral indices (steeper than 1.2). These objects represent cases of radio sources in the remnant or the restarted phases of their life cycle. Our findings confirm that these objects are not as rare as previously thought, suggesting a relatively fast cycle. They also show a variety of properties that can be relevant for modelling the evolution of radio galaxies. For example, the restarted activity can occur while the remnant structure from a previous phase of activity is still visible. This provides constraints on the duration of the “off” (dying) phase. In extended remnants with ultra-steep spectra at low frequencies, the activity likely stopped a few hundred megayears ago, and they correspond to the older tail of the age distribution of radio galaxies, in agreement with the results of simulations of radio source evolution. We find remnant radio sources with a variety of structures (from double-lobed to amorphous), possibly suggesting different types of progenitors. The present work sets the stage for exploiting the powerful tool of low-frequency spectral index studies of extended sources by taking advantage of the large areas common to the LOFAR and the Apertif surveys.

show abstract

Section: Conclusion and Future Outlookmentioning

confidence: 99%

The best of both worlds: Combining LOFAR and Apertif to derive resolved radio spectral index images

Morganti

Oosterloo

Brienza

et al. 2021

A&A

View full text Add to dashboard Cite

show abstract

“…photometric redshift estimation(Mu et al 2020), prediction of galaxy halo masses(Calderon & Berlind 2019), gravitational lenses search(Khramtsov et al 2019b), automating discovery and classification of variable stars(Bloom et al 2012), and for analyzing huge observational surveys, for example the Zwicky Transient Facility(Mahabal et al 2019), or finding planets and exocomets from the Kepler and TESS surveys(Kohler 2018).Among other recent examples, we also note the determination of physical properties of galaxies (density, metallicity, column density, ionization) from their emission-line spectra with support-vector machine algorithms employed and developed in a new GAME numerical code byUcci et al (2017); prediction of the HI content of massive galaxies at z < 2 based on optical photometry data (SDSS) and environmental parameters, which was performed byRafieferantsoa et al (2018) with regressors and deep neural network (see also examples of applications of the AstroML Python module 5 for the large-scale observational extragalactic surveys 3 ). In addition to the traditional approach for classifying the galaxy types automatically in the optical range, the machine learning methods also demonstrate a strong utility for classifying the radio galaxy types and peculiarities(Aniyan & Thorat (2017);Alger et al (2018); Wagner et al (2019);Lukic et al (2019), andRalph et al (2019)). When implementing machine learning methods for different astronomical tasks it is very useful to discuss their advantages and problem points, data quality regularity, and flexibility of the classification pipeline.…”

mentioning

confidence: 99%

Machine learning technique for morphological classification of galaxies from the SDSS

Vavilova

Dobrycheva

Vasylenko

et al. 2021

A&A

View full text Add to dashboard Cite

Context. Machine learning methods are effective tools in astronomical tasks for classifying objects by their individual features. One of the promising utilities is related to the morphological classification of galaxies at different redshifts. Aims. We use the photometry-based approach for the SDSS data 1) to exploit five supervised machine learning techniques and define the most effective among them for the automated galaxy morphological classification; 2) to test the influence of photometry data on morphology classification; 3) to discuss problem points of supervised machine learning and labeling bias; and 4) to apply the best fitting machine learning methods for revealing the unknown morphological types of galaxies from the SDSS DR9 at z < 0.1. Methods. We used different galaxy classification techniques: human labeling, multi-photometry diagrams, naive Bayes, logistic regression, support-vector machine, random forest, k-nearest neighbors. Results. We present the results of a binary automated morphological classification of galaxies conducted by human labeling, multiphotometry, and five supervised machine learning methods. We applied it to the sample of galaxies from the SDSS DR9 with redshifts of 0.02 < z < 0.1 and absolute stellar magnitudes of −24 m < M r < −19.4 m . For the analysis we used absolute magnitudes M

show abstract

“…In astrophysics, AEs have been used for feature learning in galaxy SEDs (Frontera-Pons et al 2017), image denoising (Lucas et al 2018;Ma et al 2019), and event classification (Naul et al 2018;Pasquet et al 2019). AEs are also increasingly being used in the astrophysics literature for dimensionality reduction (see, e.g., Ralph et al 2019 andPortillo et al 2020 for recent examples).…”

Section: Unsupervised Learning: a Recurrent Autoencoder Neural Network (Raenn)mentioning

confidence: 99%

SuperRAENN: A Semisupervised Supernova Photometric Classification Pipeline Trained on Pan-STARRS1 Medium-Deep Survey Supernovae

Villar

Hosseinzadeh

Berger

et al. 2020

ApJ

View full text Add to dashboard Cite

Automated classification of supernovae (SNe) based on optical photometric light-curve information is essential in the upcoming era of wide-field time domain surveys, such as the Legacy Survey of Space and Time (LSST) conducted by the Rubin Observatory. Photometric classification can enable real-time identification of interesting events for extended multiwavelength follow-up, as well as archival population studies. Here we present the complete sample of 5243 "SN-like" light curves (in g P1 r P1 i P1 z P1 ) from the Pan-STARRS1 Medium-Deep Survey (PS1-MDS). The PS1-MDS is similar to the planned LSST Wide-Fast-Deep survey in terms of cadence, filters, and depth, making this a useful training set for the community. Using this data set, we train a novel semisupervised machine learning algorithm to photometrically classify 2315 new SN-like light curves with host galaxy spectroscopic redshifts. Our algorithm consists of an RF supervised classification step and a novel unsupervised step in which we introduce a recurrent autoencoder neural network (RAENN). Our final pipeline, dubbed SuperRAENN, has an accuracy of 87% across five SN classes (Type Ia, Ibc, II, IIn, SLSN-I) and macro-averaged purity and completeness of 66% and 69%, respectively. We find the highest accuracy rates for SNe Ia and SLSNe and the lowest for SNe Ibc. Our complete spectroscopically and photometrically classified samples break down into 62.0% Type Ia (1839 objects), 19.8% Type II (553 objects), 4.8% Type IIn (136 objects), 11.7% Type Ibc (291 objects), and 1.6% Type I SLSNe (54 objects).

show abstract

Radio Galaxy Zoo: Unsupervised Clustering of Convolutionally Auto-encoded Radio-astronomical Images

Cited by 53 publications

References 33 publications

The best of both worlds: Combining LOFAR and Apertif to derive resolved radio spectral index images

The best of both worlds: Combining LOFAR and Apertif to derive resolved radio spectral index images

Machine learning technique for morphological classification of galaxies from the SDSS

SuperRAENN: A Semisupervised Supernova Photometric Classification Pipeline Trained on Pan-STARRS1 Medium-Deep Survey Supernovae

Contact Info

Product

Resources

About