Self-Supervised Representation Learning for JWST Galaxy Images and Low Resolution Spectra

Abstract

A self-supervised multimodal model recovers galaxy stellar masses from sparse JWST data that lack the multi-band coverage required by traditional SED fitting. We train on single-band F150W imaging paired with observed-frame 1-2 $\mu$m PRISM spectra and obtain a joint image-spectrum representation that recovers stellar mass at $1<z\le 3$ with $R^2\sim 0.9$ and ${\sigma }_{\mathrm{NMAD}}=0.24$ dex; cross-modal CLIP alignment further raises single-modality performance by $\Delta R^2=+0.04$ (imaging) and $+0.09$ (spectroscopy), narrowing the gap to joint-modality inference. The code and documentation are available at \href{https://github.com/2003cy/neurips2026_ssl_jwst}{GitHub}.

arXiv comment5 pages, 2 figures. Accepted for the NeurIPS 2026 RPS Workshop. Comments welcome. Detailed descriptions of the data, model training and additional benchmarks are included in the appendix (11 pages, 4 figures)