DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Learning to predict the cosmological structure formation

Journal Article · · Proceedings of the National Academy of Sciences of the United States of America
ORCiD logo [1];  [2];  [3];  [4];  [5];  [6];  [7]
  1. Carnegie Mellon Univ., Pittsburgh, PA (United States); Flatiron Inst. New York, NY (United States); Univ. of Tokyo (Japan)
  2. Univ. of California, Berkeley, CA (United States); Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States); Univ. of Tokyo (Japan)
  3. Univ. of California, Berkeley, CA (United States); Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
  4. Carnegie Mellon Univ., Pittsburgh, PA (United States); Flatiron Inst. New York, NY (United States); Univ. of Tokyo (Japan); Univ. of California, Berkeley, CA (United States); Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
  5. Univ. of British Columbia, Vancouver, BC (Canada)
  6. Flatiron Inst. New York, NY (United States)
  7. Carnegie Mellon Univ., Pittsburgh, PA (United States)

Matter evolved under the influence of gravity from minuscule density fluctuations. Nonperturbative structure formed hierarchically over all scales and developed non-Gaussian features in the Universe, known as the cosmic web. To fully understand the structure formation of the Universe is one of the holy grails of modern astrophysics. Astrophysicists survey large volumes of the Universe and use a large ensemble of computer simulations to compare with the observed data to extract the full information of our own Universe. However, to evolve billions of particles over billions of years, even with the simplest physics, is a daunting task. We build a deep neural network, the Deep Density Displacement Model ( D 3 M ), which learns from a set of prerun numerical simulations, to predict the nonlinear large-scale structure of the Universe with the Zel’dovich Approximation (ZA), an analytical approximation based on perturbation theory, as the input. Our extensive analysis demonstrates that D 3 M outperforms the second-order perturbation theory (2LPT), the commonly used fast-approximate simulation method, in predicting cosmic structure in the nonlinear regime. We also show that D 3 M is able to accurately extrapolate far beyond its training data and predict structure formation for significantly different cosmological parameters. Our study proves that deep learning is a practical and accurate alternative to approximate 3D simulations of the gravitational structure formation of the Universe.

Research Organization:
Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States)
Sponsoring Organization:
USDOE Office of Science (SC); National Aeronautics and Space Administration (NASA); Simons Foundation
Grant/Contract Number:
AC02-05CH11231
OSTI ID:
1561927
Journal Information:
Proceedings of the National Academy of Sciences of the United States of America, Journal Name: Proceedings of the National Academy of Sciences of the United States of America Journal Issue: 28 Vol. 116; ISSN 0027-8424
Publisher:
National Academy of Sciences, Washington, DC (United States)Copyright Statement
Country of Publication:
United States
Language:
English

References (85)

Lagrangian theory of gravitational instability of Friedman-Lematre cosmologies - a generic third-order model for non-linear clustering journal April 1994
Physics 101: Learning Physical Object Properties from Unlabeled Videos conference January 2016
Densely Connected Convolutional Networks preprint January 2016
Photometric Supernova Classification With Machine Learning text January 2016
The Sdss-Iv Extended Baryon Oscillation Spectroscopic Survey: Overview and Early data journal February 2016
Cosmology and Fundamental Physics with the Euclid Satellite journal September 2013
Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning journal July 2015
Exploring the posterior surface of the large scale structure reconstruction journal July 2018
Mastering the game of Go with deep neural networks and tree search journal January 2016
A New Parallel P3M Code for Very Large-Scale Cosmological Simulations text January 1998
TreePM: A code for cosmological N-body simulations journal December 2002
The Zel'dovich approximation journal February 2014
U-Net: Convolutional Networks for Biomedical Image Segmentation
  • Ronneberger, Olaf; Fischer, Philipp; Brox, Thomas
  • Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III https://doi.org/10.1007/978-3-319-24574-4_28
book November 2015
The Zeldovich approximation text January 2014
Learning Physical Intuition of Block Towers by Example preprint January 2016
Deep learning for regulatory genomics journal August 2015
Lagrangian theory of gravitational instability of Friedman-Lemaitre cosmologies - generic third-order model for non-linear clustering text January 1993
Ready or Not, Here We Go journal June 2016
Long-term Recurrent Convolutional Networks for Visual Recognition and Description preprint January 2014
Human-level control through deep reinforcement learning journal February 2015
Cosmology and fundamental physics with the Euclid satellite text January 2012
Cosmology and Fundamental Physics with the Euclid Satellite journal September 2013
Simulation as an engine of physical scene understanding journal October 2013
Cosmology and fundamental physics with the Euclid satellite text January 2018
Planck 2015 results : XXIII. The thermal Sunyaev-Zeldovich effect-cosmic infrared background correlation journal September 2016
Planck 2015 results. XIII. Cosmological parameters text January 2015
FastPM: a new scheme for fast simulations of dark matter and haloes journal August 2016
Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning journal July 2015
Planck 2015 results : XIII. Cosmological parameters journal September 2016
Planning chemical syntheses with deep neural networks and symbolic AI journal March 2018
Planning chemical syntheses with deep neural networks and symbolic AI journal March 2018
Long-term recurrent convolutional networks for visual recognition and description conference June 2015
AI designs organic syntheses journal March 2018
Simulation as an engine of physical scene understanding journal October 2013
Exploring the posterior surface of the large scale structure reconstruction journal July 2018
Learning Visual Predictive Models of Physics for Playing Billiards preprint January 2015
Bayesian physical reconstruction of initial conditions from large-scale structure surveys journal April 2013
The Baryon Oscillation Spectroscopic Survey of Sdss-Iii journal December 2012
Planck 2015 results : XXVI. The Second journal September 2016
Galaxy And Mass Assembly (GAMA): end of survey report and data release 2 journal July 2015
A volumetric deep Convolutional Neural Network for simulation of mock dark matter halo catalogues journal November 2018
DeepSphere: Efficient spherical convolutional neural network with HEALPix sampling for cosmological applications journal April 2019
Long-Term Recurrent Convolutional Networks for Visual Recognition and Description journal April 2017
Accelerating Eulerian Fluid Simulation With Convolutional Networks preprint January 2016
The Baryon Oscillation Spectroscopic Survey of SDSS-III text January 2012
Deep learning journal May 2015
A new parallel code for very large-scale cosmological simulations journal December 1998
CMU DeepLens: deep learning for automatic image-based galaxy–galaxy strong lens finding journal July 2017
CosmoFlow: Using Deep Learning to Learn the Universe at Scale
  • Mathuriya, Amrita; Bard, Deborah; Mendygral, Peter
  • SC18: International Conference for High Performance Computing, Networking, Storage and Analysis https://doi.org/10.1109/SC.2018.00068
conference November 2018
The 2dF Galaxy Redshift Survey: spectra and redshifts journal December 2001
Sdss-Iii: Massive Spectroscopic Surveys of the Distant Universe, the Milky way, and Extra-Solar Planetary Systems journal August 2011
Planck 2015 results : XVI. Isotropy and statistics of the CMB journal September 2016
Bayesian physical reconstruction of initial conditions from large-scale structure surveys journal April 2013
Cosmology and fundamental physics with the Euclid satellite journal April 2018
GADGET: a code for collisionless and gasdynamical cosmological simulations journal April 2001
Fast automated analysis of strong gravitational lenses with convolutional neural networks journal August 2017
Planck 2015 results : X. Diffuse component separation: Foreground maps journal September 2016
How filaments of galaxies are woven into the cosmic web journal April 1996
The 6dF Galaxy Survey: final redshift release (DR3) and southern large-scale structures journal October 2009
The initial conditions of the Universe from constrained simulations journal November 2012
Exploring the posterior surface of the large scale structure reconstruction text January 2018
Opportunities and obstacles for deep learning in biology and medicine journal April 2018
A volumetric deep Convolutional Neural Network for simulation of mock dark matter halo catalogues text January 2018
Machine learning for quantum physics journal February 2017
Opportunities and obstacles for deep learning in biology and medicine posted_content January 2018
The SDSS-IV extended Baryon Oscillation Spectroscopic Survey: Overview and Early Data text January 2015
V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation preprint January 2016
Fast automated analysis of strong gravitational lenses with convolutional neural networks journal August 2017
Deep learning journal May 2015
Deep Learning text January 2018
Helmholtz decomposition of the Lagrangian displacement journal April 2014
Photometric Supernova Classification with Machine Learning journal August 2016
nbodykit: An Open-source, Massively Parallel Toolkit for Large-scale Structure journal September 2018
Solving the quantum many-body problem with artificial neural networks journal February 2017
Computing the three-point correlation function of galaxies in $\mathcal {O}(N^2)$ time journal October 2015
Star–galaxy classification using deep convolutional neural networks journal October 2016
Galaxy And Mass Assembly (GAMA): end of survey report and data release 2 journal July 2015
V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation conference October 2016
The evolution of large-scale structure in a universe dominated by cold dark matter journal May 1985
Densely Connected Convolutional Networks conference July 2017
Computing the Three-Point Correlation Function of Galaxies in $\mathcal{O}(N^2)$ Time text January 2015
DeepSphere: Efficient spherical convolutional neural network with HEALPix sampling for cosmological applications journal April 2019
DeepCMB: Lensing reconstruction of the cosmic microwave background with deep neural networks journal July 2019
Newtonian Image Understanding: Unfolding the Dynamics of Objects in Static Images conference June 2016
TreePM: A code for cosmological N-body simulations journal December 2002

Cited By (10)

Learning effective physical laws for generating cosmological hydrodynamics with Lagrangian Deep Learning text January 2020
A Hybrid Deep Learning Approach to Cosmological Constraints From Galaxy Redshift Surveys text January 2019
Learning to Simulate Complex Physics with Graph Networks preprint January 2020
Fast and Accurate Non-Linear Predictions of Universes with Deep Learning preprint January 2020
Painting halos from cosmic density fields of dark matter with physically motivated neural networks journal August 2019
Painting halos from cosmic density fields of dark matter with physically motivated neural networks text January 2019
Toward Improved Predictions in Ungauged Basins: Exploiting the Power of Machine Learning journal December 2019
A Hybrid Deep Learning Approach to Cosmological Constraints from Galaxy Redshift Surveys journal February 2020
Investigating cosmological GAN emulators using latent space interpolation journal July 2021
Unveiling the predictive power of static structure in glassy systems journal April 2020

Similar Records

Non-analyticity in holographic complexity near critical points
Journal Article · 2023 · Journal of Physics. A, Mathematical and Theoretical · OSTI ID:2419759

Structure of pressure-gradient-driven current singularity in ideal magnetohydrodynamic equilibrium
Journal Article · 2023 · Plasma Physics and Controlled Fusion · OSTI ID:1926781

Scalar fields in 3D asymptotically flat higher-spin gravity
Journal Article · 2021 · Journal of Physics. A, Mathematical and Theoretical · OSTI ID:1851023