Galaxy Phase-Space and Field-Level Cosmology: The Strength of Semi-Analytic Models

Semi-analytic models are a widely used approach to simulate galaxy properties within a cosmological framework, relying on simplified yet physically motivated prescriptions. They have also proven to be an efficient alternative for generating accurate galaxy catalogs, offering a faster and less computationally expensive option compared to full hydrodynamical simulations. In this paper, we demonstrate that using only galaxy $3$D positions and radial velocities, we can train a graph neural network coupled to a moment neural network to obtain a robust machine learning based model capable of estimating the matter density parameters, $Ω_{\rm m}$, with a precision of approximately 10%. The network is trained on ($25 h^{-1}$Mpc)$^3$ volumes of galaxy catalogs from L-Galaxies and can successfully extrapolate its predictions to other semi-analytic models (GAEA, SC-SAM, and Shark) and, more remarkably, to hydrodynamical simulations (Astrid, SIMBA, IllustrisTNG, and SWIFT-EAGLE). Our results show that the network is robust to variations in astrophysical and subgrid physics, cosmological and astrophysical parameters, and the different halo-profile treatments used across simulations. This suggests that the physical relationships encoded in the phase-space of semi-analytic models are largely independent of their specific physical prescriptions, reinforcing their potential as tools for the generation of realistic mock catalogs for cosmological parameter inference.

💡 Research Summary

The paper presents a novel approach to infer the matter density parameter Ωₘ from galaxy catalogs using only three‑dimensional positions and radial velocities. The authors train a machine‑learning pipeline that couples a graph neural network (GNN) with a moment neural network (MNN). Galaxies are represented as nodes in a fully connected graph; edge features encode pairwise separations and velocity differences. The GNN performs message passing (EdgeConv‑style) to learn local relational embeddings for each galaxy, while the MNN aggregates the first and second moments of all node embeddings into a global descriptor that is fed to a fully connected regressor for Ωₘ.

Training data are drawn from the CAMELS suite, specifically 25 h⁻¹ Mpc³ volumes of galaxy catalogs generated by the semi‑analytic model L‑Galaxies. The authors use a Latin Hypercube (LH) design that varies Ωₘ∈