Neural Network Acceleration of Large-scale Structure Theory Calculations

TL;DR

This paper develops neural network surrogate models to significantly accelerate large-scale structure calculations, reducing evaluation time by a factor of 1000 while maintaining high accuracy, thereby enabling faster cosmological data analysis.

Contribution

The authors create neural network models that accurately emulate LPT predictions for matter and galaxy power spectra, drastically speeding up computations in large-scale structure analysis.

Findings

Neural network surrogates achieve ~0.1% accuracy over broad scales.

Evaluation time reduced to approximately one millisecond.

Posteriors from surrogates match those from full LPT models.

Abstract

We make use of neural networks to accelerate the calculation of power spectra required for the analysis of galaxy clustering and weak gravitational lensing data. For modern perturbation theory codes, evaluation time for a single cosmology and redshift can take on the order of two seconds. In combination with the comparable time required to compute linear predictions using a Boltzmann solver, these calculations are the bottleneck for many contemporary large-scale structure analyses. In this work, we construct neural network-based surrogate models for Lagrangian perturbation theory (LPT) predictions of matter power spectra, real and redshift space galaxy power spectra, and galaxy--matter cross power spectra that attain $\sim 0.1\%$ (at one sigma) accuracy over a broad range of scales in a $w$CDM parameter space. The neural network surrogates can be evaluated in approximately one millisecond, a factor of 1000 times faster than the full Boltzmann code and LPT computations. In a simulated full-shape redshift space galaxy power spectrum analysis, we demonstrate that the posteriors obtained using our surrogates are accurate compared to those obtained using the full LPT model. We make our surrogate models public at https://github.com/sfschen/EmulateLSS, so that others may take advantage of the speed gains they provide to enable rapid iteration on analysis settings, something that is essential in complex contemporary large-scale structure analyses.