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

Download:

Current browse context:

cond-mat.dis-nn
< prev | next >
new | recent | 1707

Change to browse by:

cond-mat
cond-mat.quant-gas

References & Citations

Bookmark

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

Condensed Matter > Disordered Systems and Neural Networks

Title: Solving the Bose-Hubbard model with machine learning

Authors: Hiroki Saito
(Submitted on 31 Jul 2017)
Abstract: Motivated by the recent successful application of artificial neural networks to quantum many-body problems [G. Carleo and M. Troyer, Science {\bf 355}, 602 (2017)], a method to calculate the ground state of the Bose-Hubbard model using a feedforward neural network is proposed. The results are in good agreement with those obtained by exact diagonalization and the Gutzwiller approximation. The method of neural-network quantum states is promising for solving quantum many-body problems of ultracold atoms in optical lattices.
Comments: 4 pages, 4 figures
Subjects: Disordered Systems and Neural Networks (cond-mat.dis-nn); Quantum Gases (cond-mat.quant-gas)
Journal reference: J. Phys. Soc. Jpn. 86, 093001 (2017)
DOI: 10.7566/JPSJ.86.093001
Cite as: arXiv:1707.09723 [cond-mat.dis-nn]
  (or arXiv:1707.09723v1 [cond-mat.dis-nn] for this version)

Submission history

From: Hiroki Saito [view email]
[v1] Mon, 31 Jul 2017 05:00:32 GMT (98kb)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.