Investigating cosmological GAN emulators using latent space interpolation

@article{Tamoinas2021InvestigatingCG,
  title={Investigating cosmological GAN emulators using latent space interpolation},
  author={Andrius Tamo{\vs}iūnas and Hans A. Winther and Kazuya Koyama and David Bacon and Robert C. Nichol and Ben Mawdsley},
  journal={Monthly Notices of the Royal Astronomical Society},
  year={2021}
}
Generative adversarial networks (GANs) have been recently applied as a novel emulation technique for large-scale structure simulations. Recent results show that GANs can be used as a fast and efficient emulator for producing novel weak lensing convergence maps as well as cosmic web data in 2D and 3D. However, like any algorithm, the GAN approach comes with a set of limitations, such as an unstable training procedure, inherent randomness of the produced outputs, and difficulties when training… 

Emulation of Cosmological Mass Maps with Conditional Generative Adversarial Networks

A novel conditional GAN model is proposed that is able to generate mass maps for any pair of matter density Ω m and matter clustering strength σ 8, parameters which have the largest impact on the evolution of structures in the Universe, for a given source galaxy redshift distribution n(z).

Can denoising diffusion probabilistic models generate realistic astrophysical fields?

This work investigates the ability of score-based generative models to generate dark matter mass density from cosmological simulations and images of interstellar dust in two astrophysical contexts, and demon-strate a proof-of-concept application of the model trained on dust in denoising dust images.

Emulating cosmological multifields with generative adversarial networks

We explore the possibility of using deep learning to generate multifield images from state-of-the-art hydrodynamic simulations of the CAMELS project. We use a generative adversarial network to

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

The CAMELS Multifield Data Set: Learning the Universe’s Fundamental Parameters with Artificial Intelligence

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.

References

SHOWING 1-10 OF 39 REFERENCES

XGBoost: A Scalable Tree Boosting System

This paper proposes a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning and provides insights on cache access patterns, data compression and sharding to build a scalable tree boosting system called XGBoost.

Encoding large scale cosmological structure with Generative Adversarial Networks

This work builds and train a GAN and uses a trained GAN to construct a simple autoencoder (AE) as a first step towards building a predictive model, and shows that the AE manages to efficiently extract information from simulation images.

Modified gravity: progresses and outlook of theories, numerical techniques and observational tests

This book represents an excellent read for the modern physicist and a great contribution to modified gravity theory and observation, exploring the latest results and most recent models, numerical s...

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

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.

Model comparison of dark energy models using deep network

This work uses a combination of a variational auto-encoder and generative adversarial network to compare different dark energy models in light of observations, yielding consistent model comparison results with those derived by the standard Bayesian method.

Dark Energy Survey Year 1 Results: Wide-field mass maps via forward fitting in harmonic space

We present new wide-field weak lensing mass maps for the Year 1 Dark Energy Survey (DES) data, generated via a forward fitting approach. This method of producing maps does not impose any prior

Generative deep fields: arbitrarily sized, random synthetic astronomical images through deep learning

Generative adversarial networks (GANs) are a class of artificial neural network that can produce realistic, but artificial, images that resemble those in a training set. In typical GAN

CosmoGAN: creating high-fidelity weak lensing convergence maps using Generative Adversarial Networks

This work applies Generative Adversarial Networks to the problem of generating weak lensing convergence maps and shows that the generator network produces maps that are described by, with high statistical confidence, the same summary statistics as the fully simulated maps.

Utilizing Minkowski functionals for image analysis: a marching square algorithm

Comparing noisy experimental image data with statistical models requires a quantitative analysis of grey-scale images beyond mean values and two-point correlations. A real-space image analysis