DeepCHART: Mapping the 3D dark matter density field from Lyα forest surveys using deep learning

Le résumé fourni par la source

ABSTRACT We present Deep learning for Cosmological Heterogeneity and Astrophysical Reconstruction via Tomography (DeepCHART), a deep learning framework designed to reconstruct the 3D dark matter density field at redshift $z = 2.5$ from Ly $\alpha$ forest spectra. Leveraging a 3D variational autoencoder with a U-Net architecture, DeepCHART performs fast field-level reconstruction, accurately capturing the non-linear gravitational dynamics and baryonic processes embedded in cosmological hydrodynamical simulations. When applied to joint data sets combining Ly $\alpha$ forest absorption and coeval galaxy positions, the reconstruction quality improves further. For current surveys, such as Subaru/PFS, CLAMATO, and LATIS, with an average transverse sightline spacing of $d_\perp = 2.4h^{-1}\mathrm{cMpc}$, DeepCHART achieves high-fidelity reconstructions over the density range $0.4\lt \Delta _{\rm DM}\lt 15$, with a voxel-wise Pearson correlation coefficient of $\rho \simeq 0.77$. These reconstructions are obtained using Ly $\alpha$ forest spectra with signal-to-noise ratios as low as 2 and instrumental resolution $R=2500$, matching Subaru/PFS specifications. For future high-density surveys enabled by instruments such as ELT/MOSAIC and WST/IFS with $d_\perp \simeq 1h^{-1}\mathrm{cMpc}$, the correlation improves to $\rho \simeq 0.90$ across a wider dynamic range ($0.25\lt \Delta _{\rm DM}\lt 40$). The framework reliably recovers the dark matter density PDF as well as the power spectrum, with only mild suppression at intermediate scales. In terms of cosmic web classification, DeepCHART successfully identifies 81 per cent of voids, 75 per cent of sheets, 63 per cent of filaments, and 43 per cent of nodes. We propose DeepCHART as a powerful and scalable framework for field-level cosmological inference, readily generalizable to other observables, and offering a robust, efficient means of maximizing the scientific return of upcoming spectroscopic surveys.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.