Beyond traditional emission-line diagnostics: Using autoencoders to uncover active galactic nuclei in DESI spectra
Abstract
Context. The growing volume of spectroscopic data in modern surveys motivates the development of data-driven approaches that complement traditional emission-line diagnostics for the identification of active galactic nuclei (AGNs).Aims. We develop and validate a machine-learning (ML) framework for galaxy and AGN classification that exploits the full optical spectrum, using unsupervised learning to capture spectral features within a semi-supervised classification framework that complements classical approaches.Methods. We adapted the SPENDER autoencoder architecture to compress galaxy spectra obtained with the Dark Energy Spectroscopic Instrument (DESI) into a low-dimensional latent space that encodes physically meaningful label-agnostic information. We introduced a quantitative classification method via a k-d tree nearest-neighbor search in latent space. The model was trained on 50 222 spectra from the DESI main survey, corresponding to the internal Guadalupe dataset, released as part of data release 1 (DR1), and restricted to z ≤ 0.5. We validated the performance using labels derived from the FastSpecFit emission line measurements defining seven galaxy classes: AGN, broad-line (BL), composite, star-forming, passive, retired, and other.Results. The model achieves a high performance in identifying AGN and BL sources, with accuracies of 0.952 and 0.965, respectively. The model reliably recovers AGN and BL sources even in spectra with a low signal-to-noise ratio and identifies AGNs missed by single-diagnostic standard methods. Our classification metrics are benchmarked against traditional diagnostics, and we show that they represent lower limits of the true model performance. The latent space learned by the autoencoder not only separates AGN and non-AGN galaxies, but also correlates with key physical properties such as stellar mass and star formation rate, demonstrating that the method captures astrophysical meaningful features.Conclusions. This work demonstrates that unsupervised spectral representation learning implemented within a semi-supervised classification framework offers a powerful and scalable pathway toward more complete and reliable AGN catalogs for current and upcoming large spectroscopic surveys.Key words: techniques: spectroscopic / galaxies: active / galaxies: general / galaxies: nuclei / galaxies: Seyfert / galaxies: statistics
Related
- Changing-look Active Galactic Nuclei from the Dark Energy Spectroscopic Instrument. VII. Testing Disk–Corona Diagnostics with eROSITA
- Evolutionary Map of the Universe (EMU): Selection of active galactic nuclei and star-formation emission with machine-learning techniques
- Changing-look Active Galactic Nuclei from the Dark Energy Spectroscopic Instrument. VII. Testing Disk-Corona Diagnostics with eROSITA
- Search for an all-sky and a Galactic Ridge diffuse neutrino emission with the first 2 years of KM3NeT/ARCA data
- Evidence for a spectral steepening of the gamma-ray emission from the Galactic Center ridge
BibTeX
@article{alcolea2026beyond,
title = {{Beyond traditional emission-line diagnostics: Using autoencoders to uncover active galactic nuclei in DESI spectra}},
author = {J.A Alcolea and M Siudek and M Eriksen and M Mezcua and R Pucha and S Juneau and S Gontcho A. Gontcho and S Panda and J Aguilar and S Ahlen and D Bianchi and A Brodzeller and D Brooks and F.J Castander and T Claybaugh and A Cuceu and A de la Macorra and B Dey and P Doel and S Ferraro and A Font-Ribera and J.E Forero-Romero and E Gaztañaga and G Gutierrez and C Hahn and H.K Herrera-Alcantar and D Joyce and R Kehoe and D Kirkby and T Kisner and A Kremin and O Lahav and C Lamman and M Landriau and L Le Guillou and A Meisner and R Miquel and J Moustakas and S Nadathur and W Percival and F Prada and I Pérez-Ràfols and G Rossi and E Sanchez and E Schlafly and D Schlegel and M Schubnell and J Silber and D Sprayberry and G Tarlé and B.A Weaver and H Zou},
journal = {Astron.Astrophys.},
year = {2026},
month = oct,
doi = {10.1051/0004-6361/202660335},
eprint = {2607.07329},
archivePrefix = {arXiv},
url = {https://hal.science/hal-05828766v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Beyond traditional emission-line diagnostics: Using autoencoders to uncover active galactic nuclei in DESI spectra
AU - J.A Alcolea
AU - M Siudek
AU - M Eriksen
AU - M Mezcua
AU - R Pucha
AU - S Juneau
AU - S Gontcho A. Gontcho
AU - S Panda
AU - J Aguilar
AU - S Ahlen
AU - D Bianchi
AU - A Brodzeller
AU - D Brooks
AU - F.J Castander
AU - T Claybaugh
AU - A Cuceu
AU - A de la Macorra
AU - B Dey
AU - P Doel
AU - S Ferraro
AU - A Font-Ribera
AU - J.E Forero-Romero
AU - E Gaztañaga
AU - G Gutierrez
AU - C Hahn
AU - H.K Herrera-Alcantar
AU - D Joyce
AU - R Kehoe
AU - D Kirkby
AU - T Kisner
AU - A Kremin
AU - O Lahav
AU - C Lamman
AU - M Landriau
AU - L Le Guillou
AU - A Meisner
AU - R Miquel
AU - J Moustakas
AU - S Nadathur
AU - W Percival
AU - F Prada
AU - I Pérez-Ràfols
AU - G Rossi
AU - E Sanchez
AU - E Schlafly
AU - D Schlegel
AU - M Schubnell
AU - J Silber
AU - D Sprayberry
AU - G Tarlé
AU - B.A Weaver
AU - H Zou
PY - 2026
DA - 2026/10/26
JO - Astron.Astrophys.
DO - 10.1051/0004-6361/202660335
UR - https://hal.science/hal-05828766v1
AB - Context. The growing volume of spectroscopic data in modern surveys motivates the development of data-driven approaches that complement traditional emission-line diagnostics for the identification of active galactic nuclei (AGNs).Aims. We develop and validate a machine-learning (ML) framework for galaxy and AGN classification that exploits the full optical spectrum, using unsupervised learning to capture spectral features within a semi-supervised classification framework that complements classical approaches.Methods. We adapted the SPENDER autoencoder architecture to compress galaxy spectra obtained with the Dark Energy Spectroscopic Instrument (DESI) into a low-dimensional latent space that encodes physically meaningful label-agnostic information. We introduced a quantitative classification method via a k-d tree nearest-neighbor search in latent space. The model was trained on 50 222 spectra from the DESI main survey, corresponding to the internal Guadalupe dataset, released as part of data release 1 (DR1), and restricted to z ≤ 0.5. We validated the performance using labels derived from the FastSpecFit emission line measurements defining seven galaxy classes: AGN, broad-line (BL), composite, star-forming, passive, retired, and other.Results. The model achieves a high performance in identifying AGN and BL sources, with accuracies of 0.952 and 0.965, respectively. The model reliably recovers AGN and BL sources even in spectra with a low signal-to-noise ratio and identifies AGNs missed by single-diagnostic standard methods. Our classification metrics are benchmarked against traditional diagnostics, and we show that they represent lower limits of the true model performance. The latent space learned by the autoencoder not only separates AGN and non-AGN galaxies, but also correlates with key physical properties such as stellar mass and star formation rate, demonstrating that the method captures astrophysical meaningful features.Conclusions. This work demonstrates that unsupervised spectral representation learning implemented within a semi-supervised classification framework offers a powerful and scalable pathway toward more complete and reliable AGN catalogs for current and upcoming large spectroscopic surveys.Key words: techniques: spectroscopic / galaxies: active / galaxies: general / galaxies: nuclei / galaxies: Seyfert / galaxies: statistics
ER -