Abstract:Self-supervised learning (SSL) applied to natural images has demonstrated a remarkable ability to learn meaningful, low-dimension representations without labels, resulting in models that are adaptable to many different tasks. Until now, applications of SSL to astronomical images have been limited to Galaxy Zoo datasets, which require a significant amount of pre-processing to prepare sparse images centered on a single galaxy. With wide-field survey instruments at the forefront of the Square Kilometer Array (SKA) era, this approach to gathering training data is impractical. We demonstrate that continuum images from surveys like the MeerKAT Galactic Cluster Legacy Survey (MGCLS) can be successfully used with SSL, without extracting single-galaxy cutouts. Using the SSL framework DINO, we experiment with various preprocessing steps, augmentations, and architectures to determine the optimal approach for this data. We train both ResNet50 and Vision Transformer (ViT) backbones. Our models match state-of-the-art results (trained on Radio Galaxy Zoo) for FRI/FRII morphology classification. Furthermore, they predict the number of compact sources via linear regression with much higher accuracy. However, fine-tuning results in similar performance between our models, the state-of-the-art, and open-source models on multi-class morphology classification. Using source-rich crops from wide-field images to train multi-purpose models is an easily scalable approach that significantly reduces data preparation time. For the tasks evaluated in this work, twenty thousand crops is sufficient training data for models that produce results similar to state-of-the-art. In the future, complex tasks like source detection and characterization, together with domain-specific tasks, ought to demonstrate the true advantages of training models with radio astronomy data over natural-image foundation models.

Semi-Supervised Rotation Measure Deconvolution and its application to MeerKAT observations of galaxy clusters

Cosmic evolution of the Faraday rotation measure in the intracluster medium of galaxy clusters

Faraday Rotation Measure Synthesis for Magnetic Fields of Galaxies

Faraday rotation as a probe of radio galaxy environment in RMHD AGN jet simulations

Probing Three-dimensional Magnetic Fields. III. Synchrotron Emission and Machine Learning

Probing Three-Dimensional Magnetic Fields: III -- Synchrotron Emission and Machine Learning

Modeling the Magnetic Field in the Galactic Disk using New Rotation Measure Observations from the Very Large Array

Disentangling the Faraday rotation sky

Self-Supervised Learning on MeerKAT Wide-Field Continuum Images

Machine Learning Approach for Estimating Magnetic Field Strength in Galaxy Clusters from Synchrotron Emission

Detectability of Galactic Faraday Rotation in Multi-wavelength CMB Observations: A Cross-Correlation Analysis of CMB and Radio Maps

Deep learning-based deconvolution for interferometric radio transient reconstruction

Wide-band Rotation Measure Synthesis

Sampling the Faraday rotation sky of TNG50: Imprint of the magnetised circumgalactic medium around Milky Way-like galaxies

Deconvolving X-ray Galaxy Cluster Spectra Using a Recurrent Inference Machine

Synchrotron Intensity Gradient Revealing Magnetic Fields in Galaxy Clusters

Classifying Complex Faraday Spectra with Convolutional Neural Networks

Probing Three-Dimensional Magnetic Fields: IV -- Synchrotron Polarization Derivative and Vision Transformer