Experimental quantum speed-up in reinforcement learning agents

Saggio, Valeria; Asenbeck, Beate E.; Hamann, Arne; Strömberg, Teodor; Schiansky, Peter; Dunjko, Vedran; Friis, Nicolai; Harris, Nicholas C.; Briegel, Hans J.; Walther, Philip

Experimental quantum speed-up in reinforcement learning agents

Files

There are no files associated with this item.

Date

2021

Authors

Saggio, Valeria

Asenbeck, Beate E.

Hamann, Arne

Strömberg, Teodor

Schiansky, Peter

Dunjko, Vedran

Friis, Nicolai

Harris, Nicholas C.

Briegel, Hans J.

Walther, Philip

DOI (citable link)

10.1038/s41586-021-03242-7

Collections

Philosophie

Publication type

Journal article

Publication status

Published

Published in

Nature ; 591 (2021), 7849. - pp. 229-233. - Springer Nature. - ISSN 0028-0836. - eISSN 1476-4687

Abstract

As the field of artificial intelligence advances, the demand for algorithms that can learn quickly and efficiently increases. An important paradigm within artificial intelligence is reinforcement learning1, where decision-making entities called agents interact with environments and learn by updating their behaviour on the basis of the obtained feedback. The crucial question for practical applications is how fast agents learn2. Although various studies have made use of quantum mechanics to speed up the agent’s decision-making process3,4, a reduction in learning time has not yet been demonstrated. Here we present a reinforcement learning experiment in which the learning process of an agent is sped up by using a quantum communication channel with the environment. We further show that combining this scenario with classical communication enables the evaluation of this improvement and allows optimal control of the learning progress. We implement this learning protocol on a compact and fully tunable integrated nanophotonic processor. The device interfaces with telecommunication-wavelength photons and features a fast active-feedback mechanism, demonstrating the agent’s systematic quantum advantage in a setup that could readily be integrated within future large-scale quantum communication networks.

Subject (DDC)

100 Philosophy

Cite This

ISO 690

SAGGIO, Valeria, Beate E. ASENBECK, Arne HAMANN, Teodor STRÖMBERG, Peter SCHIANSKY, Vedran DUNJKO, Nicolai FRIIS, Nicholas C. HARRIS, Hans J. BRIEGEL, Philip WALTHER, 2021. Experimental quantum speed-up in reinforcement learning agents. In: Nature. Springer Nature. 591(7849), pp. 229-233. ISSN 0028-0836. eISSN 1476-4687. Available under: doi: 10.1038/s41586-021-03242-7

BibTex

@article{Saggio2021-03Exper-53255,
  year={2021},
  doi={10.1038/s41586-021-03242-7},
  title={Experimental quantum speed-up in reinforcement learning agents},
  number={7849},
  volume={591},
  issn={0028-0836},
  journal={Nature},
  pages={229--233},
  author={Saggio, Valeria and Asenbeck, Beate E. and Hamann, Arne and Strömberg, Teodor and Schiansky, Peter and Dunjko, Vedran and Friis, Nicolai and Harris, Nicholas C. and Briegel, Hans J. and Walther, Philip}
}

RDF

<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/53255">
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/40"/>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/40"/>
    <dc:contributor>Harris, Nicholas C.</dc:contributor>
    <dc:language>eng</dc:language>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-03-25T08:32:09Z</dc:date>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dc:creator>Saggio, Valeria</dc:creator>
    <dc:contributor>Hamann, Arne</dc:contributor>
    <dc:creator>Hamann, Arne</dc:creator>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-03-25T08:32:09Z</dcterms:available>
    <dc:contributor>Dunjko, Vedran</dc:contributor>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/53255"/>
    <dc:creator>Briegel, Hans J.</dc:creator>
    <dc:contributor>Briegel, Hans J.</dc:contributor>
    <dc:creator>Harris, Nicholas C.</dc:creator>
    <dc:contributor>Saggio, Valeria</dc:contributor>
    <dc:creator>Dunjko, Vedran</dc:creator>
    <dc:creator>Friis, Nicolai</dc:creator>
    <dc:creator>Asenbeck, Beate E.</dc:creator>
    <dc:creator>Schiansky, Peter</dc:creator>
    <dc:contributor>Asenbeck, Beate E.</dc:contributor>
    <dcterms:issued>2021-03</dcterms:issued>
    <dc:contributor>Strömberg, Teodor</dc:contributor>
    <dc:contributor>Schiansky, Peter</dc:contributor>
    <dc:creator>Strömberg, Teodor</dc:creator>
    <dc:creator>Walther, Philip</dc:creator>
    <dc:contributor>Walther, Philip</dc:contributor>
    <dcterms:title>Experimental quantum speed-up in reinforcement learning agents</dcterms:title>
    <dc:contributor>Friis, Nicolai</dc:contributor>
    <dcterms:abstract xml:lang="eng">As the field of artificial intelligence advances, the demand for algorithms that can learn quickly and efficiently increases. An important paradigm within artificial intelligence is reinforcement learning1, where decision-making entities called agents interact with environments and learn by updating their behaviour on the basis of the obtained feedback. The crucial question for practical applications is how fast agents learn2. Although various studies have made use of quantum mechanics to speed up the agent’s decision-making process3,4, a reduction in learning time has not yet been demonstrated. Here we present a reinforcement learning experiment in which the learning process of an agent is sped up by using a quantum communication channel with the environment. We further show that combining this scenario with classical communication enables the evaluation of this improvement and allows optimal control of the learning progress. We implement this learning protocol on a compact and fully tunable integrated nanophotonic processor. The device interfaces with telecommunication-wavelength photons and features a fast active-feedback mechanism, demonstrating the agent’s systematic quantum advantage in a setup that could readily be integrated within future large-scale quantum communication networks.</dcterms:abstract>
  </rdf:Description>
</rdf:RDF>

Refereed

Yes