Cosmology with Topological Deep Learning

Abstract

The standard cosmological model with cold dark matter posits a hierarchical formation of structures. We introduce topological neural networks (TNNs), implemented as message-passing neural networks on higher-order structures, to effectively capture the topological information inherent in these hierarchies that traditional graph neural networks (GNNs) fail to account for. Our approach not only considers the vertices and edges that comprise a graph but also extends to higher-order cells such as tetrahedra, clusters, and hyperedges. This enables message-passing between these heterogeneous structures within a combinatorial complex. Furthermore, our TNNs are designed to conserve the $E(3)$-invariance, which refers to the symmetry arising from invariance against translations, reflections, and rotations. When applied to the Quijote suite, our TNNs achieve a significant reduction in the mean squared error. Compared to our GNNs, which lack higher-order message-passing, ClusterTNNs show improvements of up to 22% in ${\Omega }_{\mathrm{m}}$ and 34% in ${\sigma }_8$ jointly, while the best FullTNN achieves an improvement of up to 60% in ${\sigma }_8$. In the context of the CAMELS suite, our models yield results comparable to the current GNN benchmark, albeit with a slight decrease in performance. We emphasize that our topology and symmetry-aware neural networks provide enhanced expressive power in modeling the large-scale structures of our universe.

19 pages, 5 figures; submitted to ApJ

References in corpus (37)

• Scikit-learn: Machine Learning in Python
• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• Array Programming with NumPy
• Wilkinson Microwave Anisotropy Probe (WMAP) Three Year Results: Implications for Cosmology
• The cosmological simulation code GADGET-2
• The clustering of galaxies in the completed SDSS-III Baryon Oscillation Spectroscopic Survey: cosmological analysis of the DR12 galaxy sample
• E pur si muove: Galiliean-invariant cosmological hydrodynamical simulations on a moving mesh
• Simulating Galaxy Formation with the IllustrisTNG Model
• The Rockstar Phase-Space Temporal Halo Finder and the Velocity Offsets of Cluster Cores
• Simulating galaxy formation with black hole driven thermal and kinetic feedback
• Dark Energy Survey Year 3 Results: Cosmological Constraints from Galaxy Clustering and Weak Lensing
• DESI 2024 VI: Cosmological Constraints from the Measurements of Baryon Acoustic Oscillations
• The Arepo public code release
• Cosmological Parameters from the BOSS Galaxy Power Spectrum
• The CAMELS project: Cosmology and Astrophysics with MachinE Learning Simulations
• Cosmology and the Bispectrum
• Cosmological Constraints from Galaxy Clustering and the Mass-to-Number Ratio of Galaxy Clusters
• Towards an Optimal Estimation of Cosmological Parameters with the Wavelet Scattering Transform
• Fisher for complements: Extracting cosmology and neutrino mass from the counts-in-cells PDF
• Constraining $M_{\nu }$ with the Bispectrum II: The Total Information Content of the Galaxy Bispectrum
• New Interpretable Statistics for Large Scale Structure Analysis and Generation
• Detecting Neutrino Mass by Combining Matter Clustering, Halos, and Voids
• Cosmology with the Galaxy Bispectrum Multipoles: Optimal Estimation and Application to BOSS Data
• The CAMELS project: public data release
• Cosmological Information in the Marked Power Spectrum of the Galaxy Field
• A Hybrid Deep Learning Approach to Cosmological Constraints From Galaxy Redshift Surveys
• Quijote-PNG: Simulations of primordial non-Gaussianity and the information content of the matter field power spectrum and bispectrum
• Robust Field-level Likelihood-free Inference with Galaxies
• What does the Marked Power Spectrum Measure? Insights from Perturbation Theory
• Cluster-void degeneracy breaking: Neutrino properties and dark energy
• The Cosmic Graph: Optimal Information Extraction from Large-Scale Structure using Catalogues
• LtU-ILI: An All-in-One Framework for Implicit Inference in Astrophysics and Cosmology
• Galaxies and Halos on Graph Neural Networks: Deep Generative Modeling Scalar and Vector Quantities for Intrinsic Alignment
• Cosmology from Point Clouds with Dark Matter Halos from the Quijote Simulations
• Augmenting photometric redshift estimates using spectroscopic nearest neighbours
• Inferring Cosmological Parameters on SDSS via Domain-Generalized Neural Networks and Lightcone Simulations
• SimBIG: Field-level Simulation-Based Inference of Galaxy Clustering

Cited by in corpus (1)

• Towards an optimal extraction of cosmological parameters from galaxy cluster surveys using convolutional neural networks