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

Quantum Physics

Title: Learning Nonlinear Input-Output Maps with Dissipative Quantum Systems

Abstract: In this paper, we develop a theory of learning nonlinear input-output maps with fading memory by dissipative quantum systems, as a quantum counterpart of the theory of approximating such maps using classical dynamical systems. The theory identifies the properties required for a class of dissipative quantum systems to be {\em universal}, in that any input-output map with fading memory can be approximated arbitrarily closely by an element of this class. We then introduce an example class of dissipative quantum systems that is provably universal. Numerical experiments illustrate that with a small number of qubits, this class can achieve comparable performance to classical learning schemes with a large number of tunable parameters. Further numerical analysis suggests that the exponentially increasing Hilbert space presents a potential resource for dissipative quantum systems to surpass classical learning schemes for input-output maps.
Comments: 33 pages, 12 figures, 2 tables. Close to published version in Quantum Information Processing (this https URL)
Subjects: Quantum Physics (quant-ph); Machine Learning (cs.LG); Systems and Control (eess.SY)
Journal reference: Quantum Infor. Process., 18:198, 2019
DOI: 10.1007/s11128-019-2311-9
Cite as: arXiv:1901.01653 [quant-ph]
  (or arXiv:1901.01653v3 [quant-ph] for this version)

Submission history

From: Hendra Nurdin [view email]
[v1] Mon, 7 Jan 2019 03:12:23 GMT (392kb)
[v2] Mon, 13 May 2019 06:05:13 GMT (826kb)
[v3] Thu, 16 May 2019 13:11:57 GMT (826kb)

Link back to: arXiv, form interface, contact.