arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2610.31078

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

Yang Cheng·Morgan Fouesneau·Ivelina G. Momcheva·Tobias Buck·Shengxiu Sun
View original source ↗Read source document ↗Export metadata ↓

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 $μ$m PRISM spectra and obtain a joint image-spectrum representation that recovers stellar mass at $1<z\le3$ with $R^2 \sim 0.9$ and $σ_{\mathrm{NMAD}} = 0.24$ dex; cross-modal CLIP alignment further raises single-modality performance by $ΔR^2 = +0.04$ (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}.

Explore related subjects

astro-ph.GA

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Yang Cheng, Morgan Fouesneau, Ivelina G. Momcheva, Tobias Buck, Shengxiu Sun. 2026-10-22. Self-Supervised Representation Learning for JWST Galaxy Images and Low Resolution Spectra. https://arxiv.org/abs/2610.31078

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Environmental dependence of galaxy properties in the southern GAMA regions

Using data from the Galaxy and Mass Assembly (GAMA) survey, we investigate how galaxy properties correlate with the local environment, focusing on the two southern regions of the survey (G02 and G23) that have not previously been examined in this context. We employ two-point and marked correlation functions to quantify the environmental dependence of galaxy color, stellar mass, luminosity across the $u$, $g$, $r$, $J$, and $K$ bands, as well as star formation rate (SFR) and specific star formation rate (sSFR). We also assess the impact of redshift incompleteness and cosmic variance on these clustering measurements. Our results show that $u-r$ and $g-r$ colors are most strongly correlated with local overdensity, followed by stellar mass. The sSFR exhibits a clear inverse relationship with density of the environment, consistent with the trend observed for $u$-band luminosity, which traces young stellar populations. In contrast, galaxies brighter in the $g$, $J$, and $K$ bands preferentially inhabit denser regions. By comparing our measurements from the southern regions with those from the equatorial regions of GAMA, we find that cosmic variance does not significantly influence our conclusions. However, redshift incompleteness affects the clustering measurements, as revealed through comparisons of subsets within the G02 region. The measured correlations provide key constraints for models of galaxy assembly across mass and environment, while the environmental trends in color and near-infrared luminosity offer a means to trace stellar mass growth and quenching with redshift.

astro-ph.GA↗
arXiv2026

Dark matter density profiles of the Milky Way satellite population: reconciling simulations and observations

The cusp-core problem remains a key challenge for the $Λ$CDM model. Historically, comparing inner dark matter halo slopes ($d\logρ/d\log r$) from simulations and observations has suffered from a methodological mismatch: evaluating slopes at different radii creates apparent tensions due to inconsistent definitions rather than genuine physical differences. We rigorously compare dark matter density profiles of Milky Way (MW) dwarf satellites inferred from dynamical modelling against CDM hydrodynamical simulations. Using the NIHAO and FIRE-2 suites ($M_{\star} \sim 10^3$-$10^{11}\,\rm M_\odot$), we confront simulated profiles with dynamical inferences of 16 MW satellites across four methods, strictly within their reliability limits. Crucially, we evaluate simulated slopes at observationally accessible radii, directly comparing profiles at matched stellar mass. Both observed and simulated galaxies exhibit a mass-dependent core-formation trend: inner slopes rise from cuspy values ($\approx -1.5$) at $M_{\star} \sim 10^5\,\rm M_\odot$ to core-like values ($\approx 0$ to $-0.4$) at $M_{\star} \sim 10^8\,\rm M_\odot$. Efficient core formation begins at $M_{\star} \gtrsim 10^6\,\rm M_\odot$ for centrals, where satellites display increased scatter consistent with tidal disruption. Simulated profiles agree with observations within $\lesssim 1σ$ for most systems, outperforming a NFW profile. Furthermore, differing observational models present scatter for a given system comparable to the simulation-to-observation offset. Evaluating dark matter density slopes at fixed physical radii significantly improves agreement between hydrodynamical simulations and dynamical models, revealing no clear tension when comparing profiles at fixed stellar mass, but Willman\,1 remains an unresolved outlier.

astro-ph.GA↗
arXiv2026

Tighter Dark Matter Constraints from the Projected Mass Method: A Neural Network Enhanced Method for Galaxy Groups and Clusters

Measuring the total mass of the Milky Way and nearby galaxy groups is difficult because classical dynamical estimators rely on assumptions about satellite orbital geometry that are rarely satisfied in practice, and because only a handful of satellite galaxies are typically available as kinematic tracers. We present a new framework that corrects the well-known Projected Mass Estimator (PME) using a residual neural network trained on thousands of simulated galaxy groups from the IllustrisTNG cosmological simulation. Separate networks are trained for each satellite sample size, from as few as 5 satellites up to 50, so that the correction automatically accounts for the statistical noise that dominates when only a small number of tracers is available. In tests on simulated halos, the classical PME systematically overestimates halo masses by factors of $M_{\rm proj}/M_{\rm true} = 1.30^{+0.72}_{-0.62}$ (using the 2D distance) and $1.46^{+0.97}_{-0.72}$ (using the 3D distance), with RMSE of 0.29 and 0.32 dex respectively. The neural-network correction reduces this to $M_{\rm proj}/M_{\rm true} = 1.02^{+0.30}_{-0.26}$ with an RMSE of 0.13 dex. Applied to the Milky Way, the method yields a total mass of $M_{\rm MW} = 1.144^{+0.399}_{-0.296}\times10^{12}\,M_\odot$, with estimates based on the brightest 5-10 satellites favoring a somewhat lower range of $(0.8-0.95)\times10^{12}\,M_\odot$. The modified PME gives a tighter constraint on the virial masses and the inferred dark matter content of galaxy groups and clusters.

astro-ph.GA↗