Self-supervised contrastive learning of radio data for source detection, classification and peculiar object discovery

S. Riggi, T. Cecconello, Simone Palazzo, Andrew Hopkins, N. Gupta, Cristóbal Bordiú, A. Ingallinera, C. S. Buemi, F. Bufano, F. Cavallaro, M. D. Filipović, P. Leto, S. Loru, A. C. Ruggeri, C. Trigilio, G. Umana, Fabio Vitello

TL;DR

Researchers have developed a self-supervised "foundational model" for radio astronomy by training deep learning architectures on massive, unlabelled datasets from the ASKAP and MeerKAT telescopes. By using contrastive learning (SimCLR), the models learn to identify complex cosmic structures without human-provided labels, significantly outperforming standard models pre-trained on natural images (like ImageNet) in tasks such as galaxy classification and anomaly detection.

What Problem Does This Paper Address?

We are entering the "golden age" of radio astronomy with projects like the Square Kilometre Array (SKA) set to produce millions of images. Traditional manual cataloging is impossible at this scale, and while Deep Learning (DL) offers a solution, it faces a major bottleneck: the "labeling crisis."

Supervised DL models require thousands of high-quality, human-annotated examples to work well. In radio astronomy, labels are scarce, often ambiguous, and highly unbalanced (most sources are simple "dots," while the scientifically interesting ones are rare and complex). Furthermore, most researchers currently "fine-tune" models pre-trained on ImageNet—a dataset of cats, dogs, and cars. These models often fail to capture the subtle, diffuse, and noisy features unique to radio interferometry data.

Methods

The authors employ Self-Supervised Learning (SSL), specifically a framework called SimCLR (Simple Framework for Contrastive Learning).

Instead of learning from labels, SimCLR learns by taking an unlabelled image, creating two "augmented" versions of it (via rotation, cropping, or color jittering), and teaching a neural network (a ResNet18 backbone) that these two versions represent the same object. Simultaneously, it teaches the model to distinguish these from other images in the same batch. This forces the model to learn a "latent representation"—a mathematical summary of the image's core features.

The study explored two extraction strategies:

  1. "Hulk" mode: Blindly extracting fixed-size tiles from the sky map (unbiased but mostly empty space).
  2. "Banner" mode: Centering tiles on known radio sources (smaller but higher-quality data).

Datasets

The authors created four large-scale unlabelled datasets for pre-training, totaling over 260,000 images:

  • hulk_smgps & banner_smgps: Derived from the SARAO MeerKAT Galactic Plane Survey (SMGPS).
  • hulk_emupilot & banner_emupilot: Derived from the ASKAP Evolutionary Map of the Universe (EMU) pilot survey.

For evaluation (downstream tasks), they used:

  • Radio Galaxy Zoo (RGZ): For morphology classification.
  • Custom Annotated Datasets: For source detection and "peculiar" object discovery.

Key Insights and Ideas

  • Radio-Specific Pre-training Wins: Models pre-trained on radio data achieved 7–12% higher F1-scores in classification tasks compared to those pre-trained on ImageNet. This proves that "foundational models" built on domain-specific data are superior to general-purpose ones.
  • The "Smart Hulk" Strategy: The best results came from a two-step training process: first training on the massive, "blind" hulk dataset, then refining on the smaller, source-centered banner dataset.
  • Resilience in the Low-Label Regime: The self-supervised models were remarkably effective when only a tiny fraction (less than 10%) of labeled data was available. This is a game-changer for new surveys where human labeling hasn't even begun.
  • Feature-Based Discovery: The learned features (the "latent space") naturally clustered "interesting" objects (like complex radio galaxies) away from "ordinary" ones (compact sources), making it a powerful tool for finding rare, peculiar morphologies without knowing what to look for in advance.

Key Results

  • Classification: The radio-pre-trained models reached F1-scores of ~0.8-0.9 across various galaxy types, significantly outperforming the ImageNet baseline.
  • Source Detection: Using the learned weights in a Mask R-CNN architecture led to a 15–20% boost in detecting complex, multi-island radio sources and imaging artifacts.
  • Anomaly Detection: By applying an Isolation Forest algorithm to the learned features, the researchers successfully identified "peculiar" objects with a recall of ~60% while maintaining very low contamination from "ordinary" sources (<3%).

Limitations and Future Work

The authors acknowledge that their current evaluation datasets are still relatively small, which might limit the statistical robustness of some comparisons. They also noted a "residual dependency" on noise levels in the learned features—meaning the model sometimes groups images by their background noise rather than their astronomical content.

Future Work: The team plans to scale up to millions of images and explore Vision Transformers (ViTs), which have recently shown even greater potential in self-supervised learning than the ResNet architectures used here. They also aim to integrate these models into "astronomer-in-the-loop" systems, where the AI flags interesting candidates for human experts to review in real-time.

Summary updated

Abstract

Abstract New advancements in radio data post-processing are underway within the Square Kilometre Array (SKA) precursor community, aiming to facilitate the extraction of scientific results from survey images through a semi-automated approach. Several of these developments leverage deep learning methodologies for diverse tasks, including source detection, object or morphology classification, and anomaly detection. Despite substantial progress, the full potential of these methods often remains untapped due to challenges associated with training large supervised models, particularly in the presence of small and class-unbalanced labelled datasets. Self-supervised learning has recently established itself as a powerful methodology to deal with some of the aforementioned challenges, by directly learning a lower-dimensional representation from large samples of unlabelled data. The resulting model and data representation can then be used for data inspection and various downstream tasks if a small subset of labelled data is available. In this work, we explored contrastive learning methods to learn suitable radio data representations by training the SimCLR model on large collections of unlabelled radio images taken from the ASKAP EMU and SARAO MeerKAT GPS surveys. The resulting models were fine-tuned over smaller labelled datasets, including annotated images from various radio surveys, and evaluated on radio source detection and classification tasks. Additionally, we employed the trained self-supervised models to extract features from radio images, which were used in an unsupervised search for objects with peculiar morphology in the ASKAP EMU pilot survey data. For all considered downstream tasks, we reported the model performance metrics and discussed the benefits brought by self-supervised pre-training, paving the way for building radio foundational models in the SKA era.
BibTeX
@inproceedings{riggi2024self,
  author = {S. Riggi and T. Cecconello and Simone Palazzo and Andrew Hopkins and N. Gupta and Cristóbal Bordiú and A. Ingallinera and C. S. Buemi and F. Bufano and F. Cavallaro and M. D. Filipović and P. Leto and S. Loru and A. C. Ruggeri and C. Trigilio and G. Umana and Fabio Vitello},
  title = {{Self-supervised contrastive learning of radio data for source detection, classification and peculiar object discovery}},
  booktitle = {Publications},
  year = {2024},
  doi = {10.1017/pasa.2024.84},
}

Loading related content...

Figure