Super-resolution emulator of cosmological simulations using deep physical models

@article{KodiRamanah2020SuperresolutionEO,
  title={Super-resolution emulator of cosmological simulations using deep physical models},
  author={Doogesh Kodi Ramanah and Tom Charnock and Francisco Villaescusa-Navarro and Benjamin D. Wandelt},
  journal={Monthly Notices of the Royal Astronomical Society},
  year={2020}
}
We present an extension of our recently developed Wasserstein optimized model to emulate accurate high-resolution (HR) features from computationally cheaper low-resolution (LR) cosmological simulations. Our deep physical modelling technique relies on restricted neural networks to perform a mapping of the distribution of the LR cosmic density field to the space of the HR small-scale structures. We constrain our network using a single triplet of HR initial conditions and the corresponding LR… 

Figures from this paper

Towards Universal Cosmological Emulators with Generative Adversarial Networks
TLDR
This work trains a GAN to produce both weak lensing convergence maps and dark matter overdensity field data for multiple redshifts, cosmological parameters and modified gravity models, and applies the technique of latent space interpolation to control which outputs the algorithm produces.
AI-assisted superresolution cosmological simulations
TLDR
A deep neural network is built to enhance low-resolution dark-matter simulations, generating superresolution realizations that agree remarkably well with authentic high-resolution counterparts on their statistical properties and are orders-of-magnitude faster.
Super-resolving Dark Matter Halos using Generative Deep Learning
TLDR
This work uses a variation of U-Net with a conditional Generative Adversarial Network (GAN), generating output that visually and statistically matches the high resolution target extremely well, suggesting that the method can be used to create high resolution density output over Gpc/h box-sizes from low resolution simulations with negligible computational costs.
Cosmic voids in GAN-generated maps of large-scale structure
Dimensionless machine learning: Imposing exact units equivariance
TLDR
The approach can be used to impose units equivariance across a broad range of machine learning methods which are equivariant to rotations and other groups and discusses the in-sample and out-of-sample prediction accuracy gains one can obtain in contexts like symbolic regression and emulation, where symmetry is important.
Large-scale dark matter simulations
We review the field of collisionless numerical simulations for the large-scale structure of the Universe. We start by providing the main set of equations solved by these simulations and their
Machine Learning and Cosmology
TLDR
Current and ongoing developments relating to the application of machine learning within cosmology and a set of recommendations aimed at maximizing the scientific impact of these burgeoning tools over the coming decade through both technical development as well as the fostering of emerging communities are summarized.
The CAMELS Multifield Data Set: Learning the Universe’s Fundamental Parameters with Artificial Intelligence
TLDR
A collection of hundreds of thousands of 2D maps and 3D grids containing many different properties of cosmic gas, dark matter, and stars from more than 2000 distinct simulated universes at several cosmic times, designed to train machine-learning models.
A fast particle-mesh simulation of non-linear cosmological structure formation with massive neutrinos
Quasi-N-body simulations, such as FastPM, provide a fast way to simulate cosmological structure formation, but have yet to adequately include the effects of massive neutrinos. We present a method to
...
1
2
3
...

References

SHOWING 1-10 OF 43 REFERENCES
Learning neutrino effects in Cosmology with Convolutional Neural Networks
TLDR
This work proposes a new method, based on a deep learning network (U-Net), to quickly generate simulations with massive neutrinos from standard $\Lambda$CDM simulations without neutrino, and concludes that the method accurately reproduces the 3-dimensional spatial distribution of matter down to non-linear scales.
From Dark Matter to Galaxies with Convolutional Networks
TLDR
This paper proposes to use deep learning to establish a mapping between the 3D galaxy distribution in hydrodynamic simulations and its underlying dark matter distribution, and develops a two-phase convolutional neural network architecture to generate fast galaxy catalogues.
Systematic-free inference of the cosmic matter density field from SDSS3-BOSS data
We perform an analysis of the three-dimensional cosmic matter density field traced by galaxies of the SDSS-III/BOSS galaxy sample. The systematic-free nature of this analysis is confirmed by two
Cosmological N-body simulations: a challenge for scalable generative models
TLDR
A scalable GAN approach for training a generator of N-body three-dimensional cubes, which relies on splitting the generation of the high-dimensional data into smaller parts, and using a multi-scale approach that efficiently captures global image features that might otherwise be lost in the splitting process.
Hybrid Physical-Deep Learning Model for Astronomical Inverse Problems
We present a Bayesian machine learning architecture that combines a physically motivated parametrization and an analytic error model for the likelihood with a deep generative model providing a
HIGAN: Cosmic Neutral Hydrogen with Generative Adversarial Networks
TLDR
This work uses Wasserstein Generative Adversarial Networks (WGANs) to generate new high-resolution 3D realizations of cosmic HI at $z=5, and samples reproduce the abundance of HI across 9 orders of magnitude, from the Ly$\alpha$ forest to Damped Lyman Absorbers.
The Quijote Simulations
The Quijote simulations are a set of 43100 full N-body simulations spanning more than 7000 cosmological models in the $\{\Omega_{\rm m}, \Omega_{\rm b}, h, n_s, \sigma_8, M_\nu, w \}$ hyperplane. At
Planck 2018 results
Author(s): Aghanim, N; Akrami, Y; Ashdown, M; Aumont, J; Baccigalupi, C; Ballardini, M; Banday, AJ; Barreiro, RB; Bartolo, N; Basak, S; Battye, R; Benabed, K; Bernard, JP; Bersanelli, M; Bielewicz,
LSST: From Science Drivers to Reference Design and Anticipated Data Products
TLDR
The LSST design is driven by four main science themes: probing dark energy and dark matter, taking an inventory of the solar system, exploring the transient optical sky, and mapping the Milky Way.
Euclid Definition Study Report
Euclid is a space-based survey mission from the European Space Agency designed to understand the origin of the Universe's accelerating expansion. It will use cosmological probes to investigate the
...
1
2
3
4
5
...