Cosmological N-body simulations: a challenge for scalable generative models

Abstract

Deep generative models, such as Generative Adversarial Networks (GANs) or Variational Autoencoders (VAs) have been demonstrated to produce images of high visual quality. However, the existing hardware on which these models are trained severely limits the size of the images that can be generated. The rapid growth of high dimensional data in many fields of science therefore poses a significant challenge for generative models. In cosmology, the large-scale, three-dimensional matter distribution, modeled with N-body simulations, plays a crucial role in understanding of evolution of structures in the universe. As these simulations are computationally very expensive, GANs have recently generated interest as a possible method to emulate these datasets, but they have been, so far, mostly limited to two dimensional data. In this work, we introduce a new benchmark for the generation of three dimensional N-body simulations, in order to stimulate new ideas in the machine learning community and move closer to the practical use of generative models in cosmology. As a first benchmark result, we propose a scalable GAN approach for training a generator of N-body three-dimensional cubes. Our technique relies on two key building blocks, (i) splitting the generation of the high-dimensional data into smaller parts, and (ii) using a multi-scale approach that efficiently captures global image features that might otherwise be lost in the splitting process. We evaluate the performance of our model for the generation of N-body samples using various statistical measures commonly used in cosmology. Our results show that the proposed model produces samples of high visual quality, although the statistical analysis reveals that capturing rare features in the data poses significant problems for the generative models. We make the data, quality evaluation routines, and the proposed GAN architecture publicly available at https://github.com/nperraud/3DcosmoGAN.

Nonlinear 3D cosmic web simulation with heavy-tailed generative adversarial networks

“…If scaling up is done by stitching together subvolumes, one needs to ensure continuity at the boundary of subvolumes, and also ensure that spatial modes larger than the subvolumes are properly modeled. One implementation uses super-resolution techniques and conditional neighbor information to generate large N-body volumes [27]. As described in Sec.…”

“…This is similar in concept to [19,20], which used GANs to map dark matter density fields to corresponding halo number count maps and hydrodynamical quantities, respectively. Reference [27] proposes a super-resolution scheme for generating large-scale realizations of the matter density field hierarchically, treating the scalability problem separately from sample accuracy.…”

Investigating cosmological GAN emulators using latent space interpolation

“…In addition, the GAN approach has also been used to produce realistic cosmic microwave background temperature anisotropy 2D patches as well as deep-field astronomical images (Mishra, Reddy & Nigam 2019;Smith & Geach 2019). Finally, generating full 3D cosmic web data has been discussed in Perraudin et al (2019) and Kodi Ramanah et al (2020). The cited works show that GANs are capable of reproducing a variety of cosmological simulation outputs efficiently and with high accuracy.…”

Black Hole Weather Forecasting Using Deep Learning

“…The use of DL in cosmological simulations is a new and promising application (Perraudin et al, 2019;Mathuriya et al, 2018;He et al, 2019). Cosmological simulations are examples of simulations with high computational cost and time.…”

“…Cosmological simulations are examples of simulations with high computational cost and time. DL can potentially speedup those simulations and make it possible to investigate a larger fraction of the parameter space (Perraudin et al, 2019).…”