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

Download:

Current browse context:

hep-th

Change to browse by:

References & Citations

Bookmark

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

High Energy Physics - Theory

Title: Artificial Neural Network in Cosmic Landscape

Authors: Junyu Liu
Abstract: In this paper we propose that artificial neural network, the basis of machine learning, is useful to generate the inflationary landscape from a cosmological point of view. Traditional numerical simulations of a global cosmic landscape typically need an exponential complexity when the number of fields is large. However, a basic application of artificial neural network could solve the problem based on the universal approximation theorem of the multilayer perceptron. A toy model in inflation with multiple light fields is investigated numerically as an example of such an application.
Comments: v2, add some new contents
Subjects: High Energy Physics - Theory (hep-th); Cosmology and Nongalactic Astrophysics (astro-ph.CO); General Relativity and Quantum Cosmology (gr-qc)
Journal reference: JHEP 1712 (2017) 149
DOI: 10.1007/JHEP12(2017)149
Report number: CALT-TH-2017-048
Cite as: arXiv:1707.02800 [hep-th]
  (or arXiv:1707.02800v2 [hep-th] for this version)

Submission history

From: Junyu Liu [view email]
[v1] Mon, 10 Jul 2017 11:28:36 GMT (236kb,D)
[v2] Thu, 28 Dec 2017 03:28:43 GMT (320kb,D)

Link back to: arXiv, form interface, contact.