We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Astrophysics > Cosmology and Nongalactic Astrophysics

Title: dm2gal: Mapping Dark Matter to Galaxies with Neural Networks

Abstract: Maps of cosmic structure produced by galaxy surveys are one of the key tools for answering fundamental questions about the Universe. Accurate theoretical predictions for these quantities are needed to maximize the scientific return of these programs. Simulating the Universe by including gravity and hydrodynamics is one of the most powerful techniques to accomplish this; unfortunately, these simulations are very expensive computationally. Alternatively, gravity-only simulations are cheaper, but do not predict the locations and properties of galaxies in the cosmic web. In this work, we use convolutional neural networks to paint galaxy stellar masses on top of the dark matter field generated by gravity-only simulations. Stellar mass of galaxies are important for galaxy selection in surveys and thus an important quantity that needs to be predicted. Our model outperforms the state-of-the-art benchmark model and allows the generation of fast and accurate models of the observed galaxy distribution.
Comments: 6 pages, 1 figure, paper accepted by the NeurIPS 2020 Machine Learning and the Physical Sciences Workshop. Code available at this https URL
Subjects: Cosmology and Nongalactic Astrophysics (astro-ph.CO)
Cite as: arXiv:2012.00186 [astro-ph.CO]
  (or arXiv:2012.00186v1 [astro-ph.CO] for this version)

Submission history

From: Noah Kasmanoff [view email]
[v1] Tue, 1 Dec 2020 00:36:53 GMT (585kb,D)

Link back to: arXiv, form interface, contact.