• Corpus ID: 227126603

Deep learning insights into cosmological structure formation

  title={Deep learning insights into cosmological structure formation},
  author={Luisa Lucie-Smith and Hiranya V. Peiris and Andrew Pontzen and Brian Nord and Jeyan Thiyagalingam},
While the evolution of linear initial conditions present in the early universe into extended halos of dark matter at late times can be computed using cosmological simulations, a theoretical understanding of this complex process remains elusive. Here, we build a deep learning framework to learn this non-linear relationship, and develop techniques to physically interpret the learnt mapping. A three-dimensional convolutional neural network (CNN) is trained to predict the mass of dark matter halos… 

Figures from this paper

Discovering the building blocks of dark matter halo density profiles with neural networks

A neural network model is presented that is trained to learn the mapping from the raw density field containing each halo to the dark matter density profile, and it is shown that the model recovers the widely-used Navarro-Frenk-White profile out to the virial radius, and can additionally describe the variability in the outer profile of the halos.

Extracting cosmological parameters from N-body simulations using machine learning techniques

  • A. Lazanu
  • Computer Science, Physics
    Journal of Cosmology and Astroparticle Physics
  • 2021
It is shown that convolutional neural networks can be employed to accurately extract Ω m and σ 8 from the N-body simulations, and that these parameters can also be found from the non-linear matter power spectrum obtained from the same suite of simulations using both random forest regressors and deep neural networks.

QuasarNet: A new research platform for the data-driven investigation of black holes

QuasarNet demonstrates the power of ML, in analyzing and exploring large datasets, and also offers a unique opportunity to interrogate the theoretical assumptions that underpin accretion and feedback models.



Learning to predict the cosmological structure formation

A deep neural network is built, the Deep Density Displacement Model (D3M), 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.

Predicting dark matter halo formation in N-body simulations with deep regression networks

This work presents an innovative pathway to predict dark matter halo formation from the initial density field using a Deep Learning algorithm and shows that splitting the segmentation problem into two distinct subtasks allows for training smaller and faster networks, while the predictive power of the pipeline remains the same.

An interpretable machine-learning framework for dark matter halo formation

A generalization of the recently proposed machine-learning framework, aiming to provide new physical insights into dark matter halo formation, is presented, and the ability of the initial conditions-to-halo mass mapping learnt from one simulation to generalize to independent simulations is verified.

From Dark Matter to Galaxies with Convolutional Networks

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.

Cosmological reconstruction from galaxy light: neural network based light-matter connection

We present a method to reconstruct the initial conditions of the universe using observed galaxy positions and luminosities under the assumption that the luminosities can be calibrated with weak

From Dark Matter to Galaxies with Convolutional Neural Networks

This paper proposes a deep learning approach to map from the dark-matter-only simulation (computationally cheaper) to the galaxy distribution (from the much costlier cosmological simulation) and shows that the result outperforms a state-of-the-art model used in the astronomical community, and provides a good trade-off between computational cost and prediction accuracy.

AI-assisted superresolution cosmological simulations

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.

A Hybrid Deep Learning Approach to Cosmological Constraints from Galaxy Redshift Surveys

A deep machine learning (ML)–based technique for accurately determining σ8 and Ωm from mock 3D galaxy surveys and a hybrid approach that merges the two to combine physically motivated summary statistics with flexible CNNs is presented.

A volumetric deep Convolutional Neural Network for simulation of dark matter halo catalogues

  • P. BergerG. Stein
  • Computer Science
    Monthly Notices of the Royal Astronomical Society
  • 2018
A three-dimensional deep Convolutional Neural Network is trained to identify dark matter protohalos directly from the cosmological initial conditions to find that the network's predictions are consistent with the non-linear ellipsoidal collapse equations used explicitly by the Peak Patch algorithm.

Super-resolution emulator of cosmological simulations using deep physical models

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