Decoupling candidate dual AGN from chance superpositions in the GOTHIC survey via a deep-learning framework
Abstract
Dual active galactic nuclei (DAGN) mark a critical phase in the evolution of merging galaxies and the pairing of supermassive black holes, yet they remain difficult to identify in large imaging surveys because of projection effects and limited spatial resolution. Compact foreground stars and unresolved substructure can mimic dual nuclei through chance superposition, complicating automated detection. We revisit the 46,061 galaxies flagged but rejected as DAGN candidates by the GOTHIC pipeline, primarily because the two nuclei fell within the SDSS fibre aperture or exceeded its separation threshold. We train a supervised deep-learning framework based on the YOLOv11 oriented-bounding-box architecture on annotated SDSS imaging to separate genuine dual nuclei from foreground stellar contaminants and other spurious alignments. The final model attains a validation precision of 0.919, recall of 0.905, and F1 of 0.912, and yields 29,605 dual-nucleus candidates after removing star-dominated and blended detections. Structured visual inspection indicates that $54.5−−62\%$ are consistent with genuine dual nuclei, implying ∼(1.4--1.8)×104 plausible systems. Cross-calibrating the YOLO separation against the deterministic GOTHIC centroid measurement and restricting to the compact regime (d ≤ 6.87'') gives a conservative subset of ∼ 13,672 candidates, reaching calibrated separations of ∼ 0.56''. Spectroscopy of the most compact (≤ 1 kpc) systems shows spectra dominated by stellar absorption, with no resolved double-peaked emission, so confirmation requires higher-resolution follow-up. The catalogue is a statistically refined list of candidates, not confirmed DAGN. Nonetheless, deep-learning detection substantially reduces contamination and expands the plausible DAGN census.
Related
- Multi-Axis Selective Decoupling Framework for Free-Floating Space Manipulators
- Parameter-Decoupling-Informed Learning for Superresolution: Laboratory Validation and Galaxy Imaging
- Milky Way Near Twins (MWNeTs). I. A Hierarchical Framework for Identifying Candidate Evolutionary Counterparts of the Milky Way
- A Machine Learning Quasar Candidate Catalog with Photometric Redshifts from KiDS DR5 and AllWISE
- CausticFlow: An Efficient Machine Learning Framework Combining Neural Differential Equations and Normalizing Flows for Binary Microlensing Parameter Inference
BibTeX
@misc{mukheja2026decoupling,
title = {{Decoupling candidate dual AGN from chance superpositions in the GOTHIC survey via a deep-learning framework}},
author = {Bhavesh Mukheja and Snehanshu Saha and Anwesh Bhattacharya and Mousumi Das and Françoise Combes and Sudhanshu Barway},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2608.24164},
eprint = {2608.24164},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.24164},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Decoupling candidate dual AGN from chance superpositions in the GOTHIC survey via a deep-learning framework
AU - Bhavesh Mukheja
AU - Snehanshu Saha
AU - Anwesh Bhattacharya
AU - Mousumi Das
AU - Françoise Combes
AU - Sudhanshu Barway
PY - 2026
DA - 2026/10/23
PB - arXiv
DO - 10.48550/arxiv.2608.24164
UR - https://arxiv.org/abs/2608.24164
AB - Dual active galactic nuclei (DAGN) mark a critical phase in the evolution of merging galaxies and the pairing of supermassive black holes, yet they remain difficult to identify in large imaging surveys because of projection effects and limited spatial resolution. Compact foreground stars and unresolved substructure can mimic dual nuclei through chance superposition, complicating automated detection. We revisit the 46,061 galaxies flagged but rejected as DAGN candidates by the GOTHIC pipeline, primarily because the two nuclei fell within the SDSS fibre aperture or exceeded its separation threshold. We train a supervised deep-learning framework based on the YOLOv11 oriented-bounding-box architecture on annotated SDSS imaging to separate genuine dual nuclei from foreground stellar contaminants and other spurious alignments. The final model attains a validation precision of 0.919, recall of 0.905, and F_1 of 0.912, and yields 29,605 dual-nucleus candidates after removing star-dominated and blended detections. Structured visual inspection indicates that $54.5−−62\%$ are consistent with genuine dual nuclei, implying ∼(1.4--1.8)×10^4 plausible systems. Cross-calibrating the YOLO separation against the deterministic GOTHIC centroid measurement and restricting to the compact regime (d ≤ 6.87'') gives a conservative subset of ∼ 13,672 candidates, reaching calibrated separations of ∼ 0.56''. Spectroscopy of the most compact (≤ 1 kpc) systems shows spectra dominated by stellar absorption, with no resolved double-peaked emission, so confirmation requires higher-resolution follow-up. The catalogue is a statistically refined list of candidates, not confirmed DAGN. Nonetheless, deep-learning detection substantially reduces contamination and expands the plausible DAGN census.
ER -