From Dark Matter to Galaxies with Convolutional Neural Networks

Abstract

Cosmological simulations play an important role in the interpretation of astronomical data, in particular in comparing observed data to our theoretical expectations. However, to compare data with these simulations, the simulations in principle need to include gravity, magneto-hydrodyanmics, radiative transfer, etc. These ideal large-volume simulations (gravo-magneto-hydrodynamical) are incredibly computationally expensive which can cost tens of millions of CPU hours to run. In this paper, we propose a deep learning approach to map from the dark-matter-only simulation (computationally cheaper) to the galaxy distribution (from the much costlier cosmological simulation). The main challenge of this task is the high sparsity in the target galaxy distribution: space is mainly empty. We propose a cascade architecture composed of a classification filter followed by a regression procedure. We show that our 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.

arXiv comment5 pages, 2 figures. Accepted to the Second Workshop on Machine Learning and the Physical Sciences (NeurIPS 2019)

References in corpus (4)

• Simulating Galaxy Formation with the IllustrisTNG Model
• Going Deeper with Convolutions
• Understanding Deep Convolutional Networks
• Recurrent Residual Convolutional Neural Network based on U-Net (R2U-Net) for Medical Image Segmentation

Cited by in corpus (8)

• The CAMELS Multifield Dataset: Learning the Universe's Fundamental Parameters with Artificial Intelligence
• Multi-Epoch Machine Learning 1: Unravelling Nature vs Nurture for Galaxy Formation
• Observing relativistic features in large-scale structure surveys -- II: Doppler magnification in an ensemble of relativistic simulations
• Modeling assembly bias with machine learning and symbolic regression
• High-fidelity reproduction of central galaxy joint distributions with Neural Networks
• Field-level simulation-based inference with galaxy catalogs: the impact of systematic effects
• ScamPy -- A sub-halo clustering & abundance matching based Python interface for painting galaxies on the dark matter halo/sub-halo hierarchy
• Super-resolving Dark Matter Halos using Generative Deep Learning