Morphology of Radio Sources in Representation Space
Abstract
Understanding radio source morphologies and their classification remains challenging. We previously developed a deep clustering method based on self-supervised learning to classify a subsample of radio sources from the LOFAR Two-meter Sky Survey DR2. This yielded a labelled subset used to fine-tune an ensemble of classifiers. We aim to identify rare morphological classes in a subsample of LoTSS-DR3, which contains > 13 million sources, beyond the 12 classes previously recognised in the DR2 sample. We further aim to characterise the resulting class distribution. We applied the classifier ensemble to derive class probabilities and representations for our samples. Among DR3 sources with low class probabilities, we searched for new clusters in representation space. Moreover, we use these clusters as centroids to classify the low-probability subset. We found that 88% of the sources fall into clearly separated clusters in representation space, coinciding with the DR2 clusters. Within the remaining sources, we identified representatives of six additional morphological classes, including rare morphologies like winged and FR-ambiguous sources. The number of sources in the resulting 18 classes exhibit a strongly skewed distribution, with four dominant classes constituting 69% of the DR3 sample, while rare morphologies account for only a small fraction. The identification of additional morphological classes shows the possibility of open-set recognition with representation learning in an unexplored dataset. Unlike anomaly detection, which flags individual sources as uncommon, this approach identifies representatives of novel morphological classes, providing a framework for discovering novel source populations. The highly skewed class distribution poses a fundamental challenge for constructing balanced training datasets and highlights the need for open-set approaches in future observations.
Related
- Understanding sub-arcsecond scale radio sources and space weather with the SKA-Low using interplanetary scintillation
- Radio Galaxies as UHECR Sources in Light of Recent Microquasar Results
- Persistent Radio Sources and Multiwavelength Counterparts of Fast Radio Bursts: Magnetar Wind Nebulae and Bow Shocks in Massive Binary Systems
- Fast Radio Bursts Associated with Persistent Radio Sources Prefer Hydrogen-poor Superluminous Supernova Host Environments
- The impact of flickering variability and magnetisation on the dynamics, stability and morphology of radio-loud AGN jets
BibTeX
@misc{perez2026morphology,
title = {{Morphology of Radio Sources in Representation Space}},
author = {Nicolas Baron Perez and Marcus Brüggen and Luisa Lucie-Smith},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.03779},
eprint = {2609.03779},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.03779},
}
RIS
TY - UNPB
TI - Morphology of Radio Sources in Representation Space
AU - Nicolas Baron Perez
AU - Marcus Brüggen
AU - Luisa Lucie-Smith
PY - 2026
DA - 2026/09/09
PB - arXiv
DO - 10.48550/arxiv.2609.03779
UR - https://arxiv.org/abs/2609.03779
AB - Understanding radio source morphologies and their classification remains challenging. We previously developed a deep clustering method based on self-supervised learning to classify a subsample of radio sources from the LOFAR Two-meter Sky Survey DR2. This yielded a labelled subset used to fine-tune an ensemble of classifiers. We aim to identify rare morphological classes in a subsample of LoTSS-DR3, which contains > 13 million sources, beyond the 12 classes previously recognised in the DR2 sample. We further aim to characterise the resulting class distribution. We applied the classifier ensemble to derive class probabilities and representations for our samples. Among DR3 sources with low class probabilities, we searched for new clusters in representation space. Moreover, we use these clusters as centroids to classify the low-probability subset. We found that 88% of the sources fall into clearly separated clusters in representation space, coinciding with the DR2 clusters. Within the remaining sources, we identified representatives of six additional morphological classes, including rare morphologies like winged and FR-ambiguous sources. The number of sources in the resulting 18 classes exhibit a strongly skewed distribution, with four dominant classes constituting 69% of the DR3 sample, while rare morphologies account for only a small fraction. The identification of additional morphological classes shows the possibility of open-set recognition with representation learning in an unexplored dataset. Unlike anomaly detection, which flags individual sources as uncommon, this approach identifies representatives of novel morphological classes, providing a framework for discovering novel source populations. The highly skewed class distribution poses a fundamental challenge for constructing balanced training datasets and highlights the need for open-set approaches in future observations.
ER -