Teaching Neural Networks to Generate Fast SunyaevZel'dovich Maps
Abstract
The thermal SunyaevZel'dovich (tSZ) and the kinematic SunyaevZel'dovich (kSZ) effects trace the distribution of electron pressure and momentum in the hot universe. These observables depend on rich multiscale physics, thus, simulated maps should ideally be based on calculations that capture baryonic feedback effects such as cooling, star formation, and other complex processes. In this paper, we train deep convolutional neural networks with a UNet architecture to map from the threedimensional distribution of dark matter to electron density, momentum, and pressure at ∼100 kpc resolution. These networks are trained on a combination of the TNG300 volume and a set of cluster zoomin simulations from the IllustrisTNG project. The neural nets are able to reproduce the power spectrum, onepoint probability distribution function, bispectrum, and crosscorrelation coefficients of the simulations more accurately than the stateoftheart semianalytical models. Our approach offers a route to capture the richness of a full cosmological hydrodynamical simulation of galaxy formation with the speed of an analytical calculation.
 Publication:

The Astrophysical Journal
 Pub Date:
 October 2020
 DOI:
 10.3847/15384357/abb80f
 arXiv:
 arXiv:2007.07267
 Bibcode:
 2020ApJ...902..129T
 Keywords:

 Largescale structure of the universe;
 SunyaevZeldovich effect;
 Convolutional neural networks;
 Hydrodynamical simulations;
 902;
 1654;
 1938;
 767;
 Astrophysics  Cosmology and Nongalactic Astrophysics
 EPrint:
 21 pages, 18 figures