Euclid Quick Data Release (Q1) Exploring galaxy properties with a multi-modal foundation model
Abstract
Modern astronomical surveys, such as the Euclid mission, produce high-dimensional, multi-modal data sets that include imaging and spectroscopic information for millions of galaxies. These data serve as an ideal benchmark for large, pre-trained multi-modal models, which can leverage vast amounts of unlabelled data. In this work, we present the first exploration of Euclid data with AstroPT, an autoregressive multi-modal foundation model trained on approximately 300 000 optical and infrared Euclid images and spectral energy distributions (SEDs) from the first Euclid Quick Data Release. We compare self-supervised pre-training with baseline fully supervised training across several tasks: galaxy morphology classification; redshift estimation; similarity searches; and outlier detection. Our results show that: (a) AstroPT embeddings are highly informative, correlating with morphology and effectively isolating outliers; (b) including infrared data helps to isolate stars, but degrades the identification of edge-on galaxies, which are better captured by optical images; (c) simple fine-tuning of these embeddings for photometric redshift and stellar mass estimation outperforms a fully supervised approach, even when using only 1% of the training labels; and (d) incorporating SED data into AstroPT via a straightforward multi-modal token-chaining method improves photo-z predictions, and allow us to identify potentially more interesting anomalies (such as ringed or interacting galaxies) compared to a model pre-trained solely on imaging data.
Cite
@article{arxiv.2503.15312,
title = {Euclid Quick Data Release (Q1) Exploring galaxy properties with a multi-modal foundation model},
author = {Euclid Collaboration and M. Siudek and M. Huertas-Company and M. Smith and G. Martinez-Solaeche and F. Lanusse and S. Ho and E. Angeloudi and P. A. C. Cunha and H. Domínguez Sánchez and M. Dunn and Y. Fu and P. Iglesias-Navarro and J. Junais and J. H. Knapen and B. Laloux and M. Mezcua and W. Roster and G. Stevens and J. Vega-Ferrero and N. Aghanim and B. Altieri and A. Amara and S. Andreon and N. Auricchio and H. Aussel and C. Baccigalupi and M. Baldi and S. Bardelli and P. Battaglia and A. Biviano and A. Bonchi and E. Branchini and M. Brescia and J. Brinchmann and S. Camera and G. Cañas-Herrera and V. Capobianco and C. Carbone and J. Carretero and S. Casas and F. J. Castander and M. Castellano and G. Castignani and S. Cavuoti and K. C. Chambers and A. Cimatti and C. Colodro-Conde and G. Congedo and C. J. Conselice and L. Conversi and Y. Copin and F. Courbin and H. M. Courtois and M. Cropper and A. Da Silva and H. Degaudenzi and G. De Lucia and A. M. Di Giorgio and J. Dinis and C. Dolding and H. Dole and F. Dubath and C. A. J. Duncan and X. Dupac and S. Dusini and S. Escoffier and M. Farina and R. Farinelli and F. Faustini and S. Ferriol and F. Finelli and S. Fotopoulou and M. Frailis and E. Franceschi and S. Galeotta and K. George and B. Gillis and C. Giocoli and J. Gracia-Carpio and B. R. Granett and A. Grazian and F. Grupp and S. Gwyn and S. V. H. Haugan and W. Holmes and I. M. Hook and F. Hormuth and A. Hornstrup and K. Jahnke and M. Jhabvala and E. Keihänen and S. Kermiche and A. Kiessling and B. Kubik and M. Kümmel and M. Kunz and H. Kurki-Suonio and Q. Le Boulc'h and A. M. C. Le Brun and D. Le Mignant and S. Ligori and P. B. Lilje and V. Lindholm and I. Lloro and G. Mainetti and D. Maino and E. Maiorano and O. Mansutti and S. Marcin and O. Marggraf and M. Martinelli and N. Martinet and F. Marulli and R. Massey and S. Maurogordato and H. J. McCracken and E. Medinaceli and S. Mei and M. Melchior and Y. Mellier and M. Meneghetti and E. Merlin and G. Meylan and A. Mora and M. Moresco and L. Moscardini and R. Nakajima and C. Neissner and S. -M. Niemi and J. W. Nightingale and C. Padilla and S. Paltani and F. Pasian and K. Pedersen and W. J. Percival and V. Pettorino and S. Pires and G. Polenta and M. Poncet and L. A. Popa and L. Pozzetti and F. Raison and A. Renzi and J. Rhodes and G. Riccio and E. Romelli and M. Roncarelli and R. Saglia and Z. Sakr and A. G. Sánchez and D. Sapone and B. Sartoris and J. A. Schewtschenko and P. Schneider and T. Schrabback and M. Scodeggio and A. Secroun and G. Seidel and M. Seiffert and S. Serrano and P. Simon and C. Sirignano and G. Sirri and L. Stanco and J. Steinwagner and P. Tallada-Crespí and A. N. Taylor and I. Tereno and S. Toft and R. Toledo-Moreo and F. Torradeflot and I. Tutusaus and L. Valenziano and J. Valiviita and T. Vassallo and G. Verdoes Kleijn and A. Veropalumbo and Y. Wang and J. Weller and A. Zacchei and G. Zamorani and F. M. Zerbi and I. A. Zinchenko and E. Zucca and V. Allevato and M. Ballardini and M. Bolzonella and E. Bozzo and C. Burigana and R. Cabanac and A. Cappi and D. Di Ferdinando and J. A. Escartin Vigo and L. Gabarra and J. Martín-Fleitas and S. Matthew and N. Mauri and R. B. Metcalf and A. Pezzotta and M. Pöntinen and C. Porciani and I. Risso and V. Scottez and M. Sereno and M. Tenti and M. Viel and M. Wiesmann and Y. Akrami and I. T. Andika and S. Anselmi and M. Archidiacono and F. Atrio-Barandela and C. Benoist and K. Benson and D. Bertacca and M. Bethermin and L. Bisigello and A. Blanchard and L. Blot and M. L. Brown and S. Bruton and A. Calabro and B. Camacho Quevedo and F. Caro and C. S. Carvalho and T. Castro and Y. Charles and F. Cogato and A. R. Cooray and O. Cucciati and S. Davini and F. De Paolis and G. Desprez and A. Díaz-Sánchez and J. J. Diaz and S. Di Domizio and J. M. Diego and P. -A. Duc and A. Enia and Y. Fang and A. G. Ferrari and P. G. Ferreira and A. Finoguenov and A. Fontana and A. Franco and K. Ganga and J. García-Bellido and T. Gasparetto and V. Gautard and E. Gaztanaga and F. Giacomini and F. Gianotti and G. Gozaliasl and M. Guidi and C. M. Gutierrez and A. Hall and W. G. Hartley and S. Hemmati and C. Hernández-Monteagudo and H. Hildebrandt and J. Hjorth and J. J. E. Kajava and Y. Kang and V. Kansal and D. Karagiannis and K. Kiiveri and C. C. Kirkpatrick and S. Kruk and J. Le Graet and L. Legrand and M. Lembo and F. Lepori and G. Leroy and G. F. Lesci and J. Lesgourgues and L. Leuzzi and T. I. Liaudat and A. Loureiro and J. Macias-Perez and G. Maggio and M. Magliocchetti and E. A. Magnier and F. Mannucci and R. Maoli and C. J. A. P. Martins and L. Maurin and M. Miluzio and P. Monaco and C. Moretti and G. Morgante and C. Murray and K. Naidoo and A. Navarro-Alsina and S. Nesseris and F. Passalacqua and K. Paterson and L. Patrizii and A. Pisani and D. Potter and S. Quai and M. Radovich and S. Sacquegna and M. Sahlén and D. B. Sanders and E. Sarpa and A. Schneider and D. Sciotti and D. Scognamiglio and E. Sellentin and L. C. Smith and K. Tanidis and G. Testera and R. Teyssier and S. Tosi and A. Troja and M. Tucci and C. Valieri and A. Venhola and D. Vergani and G. Verza and P. Vielzeuf and N. A. Walton and J. G. Sorce},
journal= {arXiv preprint arXiv:2503.15312},
year = {2025}
}
Comments
Paper submitted as part of the A&A Special Issue `Euclid Quick Data Release (Q1)', 31 pages, 17 figures