Inferring redshift and galaxy properties via a multi-task neural net with probabilistic outputs: An application to simulated MOONS spectra

Abstract

The era of large-scale astronomical surveys demands innovative approaches for rapid and accurate analysis of extensive spectral data, and a promising direction in which to address this challenge is offered by machine learning. Here, we introduce a new pipeline, M-TOPnet (Multi-Task network Outputting Probabilities), which employs a convolutional neural network with residual learning to simultaneously derive redshift and other key physical properties of galaxies from their spectra. Our tool efficiently encodes spectral information into a latent space, employing distinct downstream branches for each physical quantity, thereby benefiting from multi-task learning. Notably, our method handles the redshift output as a probability distribution, allowing for a more refined and robust estimation of this critical parameter. We demonstrate preliminary results using simulated data from the MOONS instrument, which will soon be operating at the ESO/VLT. We highlight the effectiveness of our tool in accurately predicting the redshift, stellar mass, and star formation rate of galaxies at z>~1-3, even for faint sources (m_H ~ 24) for which traditional methods often struggle. Through analysis of the output probability distributions, we demonstrate that our pipeline enables robust quality screening of the results, achieving accuracy rates of up to 99% in redshift determination (defined as predictions within |Delta_z| < 0.01 relative to the true redshift) with 8h exposure spectra, while automatically identifying potentially problematic cases. Our pipeline thus emerges as a powerful solution for the upcoming challenges in observational astronomy, combining precision, interpretability, and efficiency, all aspects that are crucial for analysing the massive datasets expected from next-generation instruments.

accepted by A&A

References in corpus (44)

• Cosmic Star Formation History
• Mass and environment as drivers of galaxy evolution in SDSS and zCOSMOS and the origin of the Schechter function
• A Highly Consistent Framework for the Evolution of the Star-Forming "Main Sequence" from z~0-6
• 3D-HST+CANDELS: The Evolution of the Galaxy Size-Mass Distribution since $z=3$
• Detection of quiescent galaxies in a bicolor sequence from z=0-2
• Constraining the Low-Mass Slope of the Star Formation Sequence at 0.5<z<2.5
• The Herschel view of the dominant mode of galaxy growth from z=4 to the present day
• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• Strong Nebular Line Ratios in the Spectra of z~2-3 Star-forming Galaxies: First Results from KBSS-MOSFIRE
• Strangulation as the primary mechanism for shutting down star formation in galaxies
• Understanding Galaxy Evolution through Emission Lines
• An Objective Definition for the Main Sequence of Star-Forming Galaxies
• The Mass-Metallicity and the Fundamental Metallicity Relation revisited on a fully Te-based abundance scale for galaxies
• The galaxy stellar mass function at 3.5<z<7.5 in the CANDELS/UDS, GOODS-South, and HUDF fields
• The chemical enrichment in the early Universe as probed by JWST via direct metallicity measurements at z~8
• DESI Bright Galaxy Survey: Final Target Selection, Design, and Validation
• Recovering galaxy star formation and metallicity histories from spectra using VESPA
• The MOSDEF Survey: Excitation Properties of $z\sim 2.3$ Star-forming Galaxies
• GOODS-HERSCHEL: star formation, dust attenuation and the FIR-radio correlation on the Main Sequence of star-forming galaxies up to z~4
• EIGER II. first spectroscopic characterisation of the young stars and ionised gas associated with strong H$\beta$ and [OIII] line-emission in galaxies at z=5-7 with JWST
• Hydrodynamical simulations of the galaxy population: enduring successes and outstanding challenges
• Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing
• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• Galaxy Merger Rates up to z $\sim$ 3 using a Bayesian Deep Learning Model $-$ A Major-Merger classifier using IllustrisTNG Simulation data
• Convolutional neural network identification of galaxy post-mergers in UNIONS using IllustrisTNG
• Astronomia ex machina: a history, primer, and outlook on neural networks in astronomy
• The mass-metallicity relations for gas and stars in star-forming galaxies: strong outflow vs variable IMF
• Where does galactic dust come from?
• Galaxy Morphology Network: A Convolutional Neural Network Used to Study Morphology and Quenching in $\sim 100,000$ SDSS and $\sim 20,000$ CANDELS Galaxies
• The assembly of dusty galaxies at $z\ge 4$: the build-up of stellar mass and its scaling relations with hints from early JWST data
• Redshift Measurement and Spectral Classification for eBOSS Galaxies with the Redmonster Software
• Spectroscopic and Photometric Redshift Estimation by Neural Networks For the China Space Station Optical Survey (CSS-OS)
• Autoencoding Galaxy Spectra I: Architecture
• Detecting and Characterizing Mg II absorption in DESI Survey Validation Quasar Spectra
• Inferring physical properties of galaxies from their emission line spectra
• Autoencoding Galaxy Spectra II: Redshift Invariance and Outlier Detection
• On the origin of star formation quenching in massive galaxies at $z\gtrsim 3$ in the cosmological simulations IllustrisTNG
• The Gaia-ESO Survey: Chemical evolution of Mg and Al in the Milky Way with Machine-Learning
• Deriving the star formation histories of galaxies from spectra with simulation-based inference
• AutoSourceID-Classifier. Star-Galaxy Classification using a Convolutional Neural Network with Spatial Information
• MusE GAs FLOw and Wind (MEGAFLOW) IX. The impact of gas flows on the relations between the mass, star formation rate and metallicity of galaxies
• Galaxy Spectra neural Network (GaSNet). II. Using Deep Learning for Spectral Classification and Redshift Predictions
• MAMBO -- An empirical galaxy and AGN mock catalogue for the exploitation of future surveys
• FNet II: Spectral Classification of Quasars, Galaxies, Stars, and broad absorption line (BAL) Quasars

Cited by in corpus (5)

• Machine learning the gap between real and simulated nebulae: A domain-adaptation approach to classify ionised nebulae in nearby galaxies
• Galaxy Spectra Networks (GaSNet). III. Generative pre-trained network for spectrum reconstruction, redshift estimate and anomaly detection
• Automated quasar continuum estimation using neural networks: a comparative study of deep-learning architectures
• Classifying spectra of emission-line regions with neural networks -- An application to integral field spectroscopic data of M33
• Accurate spectroscopic redshift estimation using non-negative matrix factorization: application to MUSE spectra