Cons-training tensor networks: Embedding and optimization over discrete linear constraints

In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary discrete linear constraints, including inequalities, into sparse block structures. These tensor networks are particularly tailored for modeling dis...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:SciPost physics Jg. 18; H. 6; S. 192
Hauptverfasser: Lopez Piqueres, Javier, Chen, Jing
Format: Journal Article
Sprache:Englisch
Veröffentlicht: SciPost 01.06.2025
ISSN:2542-4653, 2542-4653
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary discrete linear constraints, including inequalities, into sparse block structures. These tensor networks are particularly tailored for modeling distributions with support strictly over the feasible space, offering benefits such as reducing the search space in optimization problems, alleviating overfitting, improving training efficiency, and decreasing model size. Central to our approach is the concept of a quantum region, an extension of quantum numbers traditionally used in U(1) U ( 1 ) symmetric tensor networks, adapted to capture any linear constraint, including the unconstrained scenario. We further develop a novel canonical form for these new MPS, which allow for the merging and factorization of tensor blocks according to quantum region fusion rules and permit optimal truncation schemes. Utilizing this canonical form, we apply an unsupervised training strategy to optimize arbitrary objective functions subject to discrete linear constraints. Our method’s efficacy is demonstrated by solving the quadratic knapsack problem, achieving superior performance compared to a leading nonlinear integer programming solver. Additionally, we analyze the complexity and scalability of our approach, demonstrating its potential in addressing complex constrained combinatorial optimization problems.
AbstractList In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary discrete linear constraints, including inequalities, into sparse block structures. These tensor networks are particularly tailored for modeling distributions with support strictly over the feasible space, offering benefits such as reducing the search space in optimization problems, alleviating overfitting, improving training efficiency, and decreasing model size. Central to our approach is the concept of a quantum region, an extension of quantum numbers traditionally used in U(1) U ( 1 ) symmetric tensor networks, adapted to capture any linear constraint, including the unconstrained scenario. We further develop a novel canonical form for these new MPS, which allow for the merging and factorization of tensor blocks according to quantum region fusion rules and permit optimal truncation schemes. Utilizing this canonical form, we apply an unsupervised training strategy to optimize arbitrary objective functions subject to discrete linear constraints. Our method’s efficacy is demonstrated by solving the quadratic knapsack problem, achieving superior performance compared to a leading nonlinear integer programming solver. Additionally, we analyze the complexity and scalability of our approach, demonstrating its potential in addressing complex constrained combinatorial optimization problems.
In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary discrete linear constraints, including inequalities, into sparse block structures. These tensor networks are particularly tailored for modeling distributions with support strictly over the feasible space, offering benefits such as reducing the search space in optimization problems, alleviating overfitting, improving training efficiency, and decreasing model size. Central to our approach is the concept of a quantum region, an extension of quantum numbers traditionally used in $U(1)$ symmetric tensor networks, adapted to capture any linear constraint, including the unconstrained scenario. We further develop a novel canonical form for these new MPS, which allow for the merging and factorization of tensor blocks according to quantum region fusion rules and permit optimal truncation schemes. Utilizing this canonical form, we apply an unsupervised training strategy to optimize arbitrary objective functions subject to discrete linear constraints. Our method's efficacy is demonstrated by solving the quadratic knapsack problem, achieving superior performance compared to a leading nonlinear integer programming solver. Additionally, we analyze the complexity and scalability of our approach, demonstrating its potential in addressing complex constrained combinatorial optimization problems.
ArticleNumber 192
Author Lopez Piqueres, Javier
Chen, Jing
Author_xml – sequence: 1
  givenname: Javier
  surname: Lopez Piqueres
  fullname: Lopez Piqueres, Javier
– sequence: 2
  givenname: Jing
  orcidid: 0000-0003-0538-689X
  surname: Chen
  fullname: Chen, Jing
BookMark eNpN0M1KAzEUBeAgFay1jyDMC0ydO_lpxp2UqoWCBbtyEzLJnZraJiUJSn16ayvS1bncA9_iXJOeDx4JuYVqVAMT8u7VuEVIefG-TyOQIzGCpr4g_ZqzumSC097ZfUWGKa2rqqoBGhC8T94mwacyR-2886sio08hFh7zV4gf6b6Yblu09rfS3hZhl93Wfevsgi_CJ8bCumQiZiw2zqOOhTlwRy2nG3LZ6U3C4V8OyPJxupw8l_OXp9nkYV4aCiKXQrSWA2oJVvLxmPJONGCRcqSWdWPZIj8kQyGotQ3Xopa1Rqsbw8YtcDogsxNrg16rXXRbHfcqaKeOjxBXSsfszAaVoFoDFYCUamZsLVvOjKHUdBKEYN3B4ifLxJBSxO7fg0od51ZncyuQSqjD3PQH2dZ6eQ
Cites_doi 10.1007/s10898-023-01345-1
10.1007/s10288-019-00424-y
10.21468/SciPostPhysCodeb.4
10.21468/SciPostPhys.7.5.060
10.48550/arXiv.2006.02516
10.1038/s41467-024-46959-5
10.1088/1742-5468/2013/04/P04008
10.1016/j.aop.2010.09.012
10.3389/fphy.2022.906590
10.1016/j.aop.2012.07.009
10.1103/PhysRevB.86.195114
10.1088/2058-9565/aaba1a
10.1103/PhysRevX.8.031012
10.1137/22M1501787
10.1103/PhysRevB.73.085115
10.48550/arXiv.cond-mat/0407066
10.1007/s10955-015-1276-z
10.21468/SciPostPhysLectNotes.8
10.1038/s42256-021-00401-3
10.1088/2632-2153/ace0f5
10.22331/q-2024-07-25-1425
10.1103/PhysRevB.103.125117
10.1103/PhysRevLett.126.090506
10.26421/QIC8.6-7-6
10.1103/PhysRevB.83.115125
10.1088/1742-5468/2007/08/P08024
10.1103/PhysRevA.74.022320
10.21468/SciPostPhysCodeb.4-r0.3
10.48550/arXiv.2206.03832
10.1103/PhysRevB.85.165146
10.1007/s12532-023-00239-3
10.1088/1367-2630/ab31ef
ContentType Journal Article
DBID AAYXX
CITATION
DOA
DOI 10.21468/SciPostPhys.18.6.192
DatabaseName CrossRef
DOAJ Directory of Open Access Journals
DatabaseTitle CrossRef
DatabaseTitleList CrossRef

Database_xml – sequence: 1
  dbid: DOA
  name: DOAJ Directory of Open Access Journals
  url: https://www.doaj.org/
  sourceTypes: Open Website
DeliveryMethod fulltext_linktorsrc
Discipline Physics
EISSN 2542-4653
ExternalDocumentID oai_doaj_org_article_63aa1361e33a4cd28b54cc33cf81664f
10_21468_SciPostPhys_18_6_192
GroupedDBID 5VS
AAFWJ
AAYXX
ADBBV
AFPKN
ALMA_UNASSIGNED_HOLDINGS
BCNDV
CITATION
GROUPED_DOAJ
M~E
OK1
ID FETCH-LOGICAL-c316t-66bd51ea81d857735f691de35e3d4f78be5d4f4e663dd95a6282aeda9c47b153
IEDL.DBID DOA
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001511976700001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 2542-4653
IngestDate Fri Oct 03 12:44:37 EDT 2025
Sat Nov 29 07:48:16 EST 2025
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 6
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c316t-66bd51ea81d857735f691de35e3d4f78be5d4f4e663dd95a6282aeda9c47b153
ORCID 0000-0003-0538-689X
OpenAccessLink https://doaj.org/article/63aa1361e33a4cd28b54cc33cf81664f
ParticipantIDs doaj_primary_oai_doaj_org_article_63aa1361e33a4cd28b54cc33cf81664f
crossref_primary_10_21468_SciPostPhys_18_6_192
PublicationCentury 2000
PublicationDate 2025-06-01
PublicationDateYYYYMMDD 2025-06-01
PublicationDate_xml – month: 06
  year: 2025
  text: 2025-06-01
  day: 01
PublicationDecade 2020
PublicationTitle SciPost physics
PublicationYear 2025
Publisher SciPost
Publisher_xml – name: SciPost
References ref13
ref12
ref15
ref14
ref31
ref30
ref11
ref10
ref32
ref2
ref1
ref17
ref16
ref19
ref18
ref24
ref23
ref26
ref25
ref20
ref22
ref21
ref28
ref27
ref29
ref8
ref7
ref9
ref4
ref3
ref6
ref5
References_xml – ident: ref23
  doi: 10.1007/s10898-023-01345-1
– ident: ref12
  doi: 10.1007/s10288-019-00424-y
– ident: ref25
  doi: 10.21468/SciPostPhysCodeb.4
– ident: ref28
  doi: 10.21468/SciPostPhys.7.5.060
– ident: ref18
  doi: 10.48550/arXiv.2006.02516
– ident: ref20
  doi: 10.1038/s41467-024-46959-5
– ident: ref13
  doi: 10.1088/1742-5468/2013/04/P04008
– ident: ref11
  doi: 10.1016/j.aop.2010.09.012
– ident: ref30
  doi: 10.3389/fphy.2022.906590
– ident: ref8
  doi: 10.1016/j.aop.2012.07.009
– ident: ref7
  doi: 10.1103/PhysRevB.86.195114
– ident: ref15
  doi: 10.1088/2058-9565/aaba1a
– ident: ref14
  doi: 10.1103/PhysRevX.8.031012
– ident: ref32
  doi: 10.1137/22M1501787
– ident: ref1
  doi: 10.1103/PhysRevB.73.085115
– ident: ref3
  doi: 10.48550/arXiv.cond-mat/0407066
– ident: ref27
  doi: 10.1007/s10955-015-1276-z
– ident: ref9
  doi: 10.21468/SciPostPhysLectNotes.8
– ident: ref21
  doi: 10.1038/s42256-021-00401-3
– ident: ref10
  doi: 10.1088/2632-2153/ace0f5
– ident: ref19
  doi: 10.22331/q-2024-07-25-1425
– ident: ref17
  doi: 10.1103/PhysRevB.103.125117
– ident: ref29
  doi: 10.1103/PhysRevLett.126.090506
– ident: ref4
  doi: 10.26421/QIC8.6-7-6
– ident: ref6
  doi: 10.1103/PhysRevB.83.115125
– ident: ref2
  doi: 10.1088/1742-5468/2007/08/P08024
– ident: ref5
  doi: 10.1103/PhysRevA.74.022320
– ident: ref26
  doi: 10.21468/SciPostPhysCodeb.4-r0.3
– ident: ref31
  doi: 10.48550/arXiv.2206.03832
– ident: ref22
  doi: 10.1103/PhysRevB.85.165146
– ident: ref24
  doi: 10.1007/s12532-023-00239-3
– ident: ref16
  doi: 10.1088/1367-2630/ab31ef
SSID ssj0002119165
Score 2.2934155
Snippet In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary...
SourceID doaj
crossref
SourceType Open Website
Index Database
StartPage 192
Title Cons-training tensor networks: Embedding and optimization over discrete linear constraints
URI https://doaj.org/article/63aa1361e33a4cd28b54cc33cf81664f
Volume 18
WOSCitedRecordID wos001511976700001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVAON
  databaseName: DOAJ Directory of Open Access Journals
  customDbUrl:
  eissn: 2542-4653
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0002119165
  issn: 2542-4653
  databaseCode: DOA
  dateStart: 20160101
  isFulltext: true
  titleUrlDefault: https://www.doaj.org/
  providerName: Directory of Open Access Journals
– providerCode: PRVHPJ
  databaseName: ROAD: Directory of Open Access Scholarly Resources
  customDbUrl:
  eissn: 2542-4653
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0002119165
  issn: 2542-4653
  databaseCode: M~E
  dateStart: 20160101
  isFulltext: true
  titleUrlDefault: https://road.issn.org
  providerName: ISSN International Centre
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV07T8MwELYQAokF8RTlJQ-saeX4EYcNUCsGqBg6VCyRY58lhqYoKYz8du6SFrKxsCRSFFnRd8ndd_HdfYzdpKAgmCgTa2xMkFJD4kDrxEgoHVDI8e0Q16dsOrXzef7Sk_qimrBuPHAH3MhI54Q0AqR0yofUllp5L6WPtOOlInlfZD29ZIp8cDu3zOiuZYe0q-0IvxXSv6XCyqGwQzPstj9_g1FvZn8bXCYHbH_NCvld9zSHbAuqI7bbVmf65pi9kqxmslFz4FR0vqx51VVwN7d8vCghUBDirgp8iV5gsW6v5FShyan1tkZ2zIlTupp7IoW02qo5YbPJePbwmKxFERIvhVklxpRBC3DIM63OMqmjyUUAqUEGFTNbgsazAmQSIeTaGcypHASXe5WV6N5O2Xa1rOCM8TRqExwogTkThjJlnYxeO6l8RF6QiQEbbsAp3rvRFwWmDC2aRQ_NQtjCFIjmgN0ThD830-Tq9gLas1jbs_jLnuf_scgF20tJp7f9W3LJtlf1B1yxHf-5emvq6_ZVwePz1_gb35XHyw
linkProvider Directory of Open Access Journals
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Cons-training+tensor+networks%3A+Embedding+and+optimization+over+discrete+linear+constraints&rft.jtitle=SciPost+physics&rft.au=Lopez+Piqueres%2C+Javier&rft.au=Chen%2C+Jing&rft.date=2025-06-01&rft.issn=2542-4653&rft.eissn=2542-4653&rft.volume=18&rft.issue=6&rft_id=info:doi/10.21468%2FSciPostPhys.18.6.192&rft.externalDBID=n%2Fa&rft.externalDocID=10_21468_SciPostPhys_18_6_192
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2542-4653&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2542-4653&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2542-4653&client=summon