Efficient PINNs via multi-head unimodular regularization of the solutions space

Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle stiff differential equations. Here we present a machine learning framework to facilitate the solution of nonlinear multiscale differential eq...

Full description

Saved in:

Bibliographic Details
Published in:	Communications physics Vol. 8; no. 1; pp. 335 - 14
Main Authors:	Tarancón-Álvarez, Pedro, Tejerina-Pérez, Pablo, Jimenez, Raul, Protopapas, Pavlos
Format:	Journal Article
Language:	English
Published:	London Nature Publishing Group UK 15.08.2025 Nature Publishing Group Nature Portfolio
Subjects:	639/705/1042 639/766/259 Accuracy Algorithms Boundary conditions Einstein equations Geometry Inverse problems Linear equations Machine learning Neural networks Nonlinear differential equations Numerical analysis Ordinary differential equations Physics Physics and Astronomy Regularization
ISSN:	2399-3650, 2399-3650
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle stiff differential equations. Here we present a machine learning framework to facilitate the solution of nonlinear multiscale differential equations and, especially, inverse problems using Physics-Informed Neural Networks (PINNs). This framework is based on what is called multi-head (MH) training, which involves training the network to learn a general space of all solutions for a given set of equations with certain variability, rather than learning a specific solution of the system. This setup is used with a second novel technique that we call Unimodular Regularization (UR) of the latent space of solutions. We show that the multi-head approach, combined with Unimodular Regularization, significantly improves the efficiency of PINNs by facilitating the transfer learning process thereby enabling the finding of solutions for nonlinear, coupled, and multiscale differential equations. Physics-Informed Neural Networks (PINNs) face challenges in generalizing solutions for nonlinear multiscale differential equations and inverse problems. Here, the authors employ a framework of multihead training with unimodular regularization, to enhance PINN efficiency and enable effective transfer learning in case of systems, including the flame equation, van der Pol oscillator, and Einstein Field Equations.
AbstractList	Abstract Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle stiff differential equations. Here we present a machine learning framework to facilitate the solution of nonlinear multiscale differential equations and, especially, inverse problems using Physics-Informed Neural Networks (PINNs). This framework is based on what is called multi-head (MH) training, which involves training the network to learn a general space of all solutions for a given set of equations with certain variability, rather than learning a specific solution of the system. This setup is used with a second novel technique that we call Unimodular Regularization (UR) of the latent space of solutions. We show that the multi-head approach, combined with Unimodular Regularization, significantly improves the efficiency of PINNs by facilitating the transfer learning process thereby enabling the finding of solutions for nonlinear, coupled, and multiscale differential equations. Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle stiff differential equations. Here we present a machine learning framework to facilitate the solution of nonlinear multiscale differential equations and, especially, inverse problems using Physics-Informed Neural Networks (PINNs). This framework is based on what is called multi-head (MH) training, which involves training the network to learn a general space of all solutions for a given set of equations with certain variability, rather than learning a specific solution of the system. This setup is used with a second novel technique that we call Unimodular Regularization (UR) of the latent space of solutions. We show that the multi-head approach, combined with Unimodular Regularization, significantly improves the efficiency of PINNs by facilitating the transfer learning process thereby enabling the finding of solutions for nonlinear, coupled, and multiscale differential equations.Physics-Informed Neural Networks (PINNs) face challenges in generalizing solutions for nonlinear multiscale differential equations and inverse problems. Here, the authors employ a framework of multihead training with unimodular regularization, to enhance PINN efficiency and enable effective transfer learning in case of systems, including the flame equation, van der Pol oscillator, and Einstein Field Equations. Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle stiff differential equations. Here we present a machine learning framework to facilitate the solution of nonlinear multiscale differential equations and, especially, inverse problems using Physics-Informed Neural Networks (PINNs). This framework is based on what is called multi-head (MH) training, which involves training the network to learn a general space of all solutions for a given set of equations with certain variability, rather than learning a specific solution of the system. This setup is used with a second novel technique that we call Unimodular Regularization (UR) of the latent space of solutions. We show that the multi-head approach, combined with Unimodular Regularization, significantly improves the efficiency of PINNs by facilitating the transfer learning process thereby enabling the finding of solutions for nonlinear, coupled, and multiscale differential equations. Physics-Informed Neural Networks (PINNs) face challenges in generalizing solutions for nonlinear multiscale differential equations and inverse problems. Here, the authors employ a framework of multihead training with unimodular regularization, to enhance PINN efficiency and enable effective transfer learning in case of systems, including the flame equation, van der Pol oscillator, and Einstein Field Equations. Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle stiff differential equations. Here we present a machine learning framework to facilitate the solution of nonlinear multiscale differential equations and, especially, inverse problems using Physics-Informed Neural Networks (PINNs). This framework is based on what is called multi-head (MH) training, which involves training the network to learn a general space of all solutions for a given set of equations with certain variability, rather than learning a specific solution of the system. This setup is used with a second novel technique that we call Unimodular Regularization (UR) of the latent space of solutions. We show that the multi-head approach, combined with Unimodular Regularization, significantly improves the efficiency of PINNs by facilitating the transfer learning process thereby enabling the finding of solutions for nonlinear, coupled, and multiscale differential equations.
ArticleNumber	335
Author	Protopapas, Pavlos Jimenez, Raul Tarancón-Álvarez, Pedro Tejerina-Pérez, Pablo
Author_xml	– sequence: 1 givenname: Pedro orcidid: 0009-0004-5774-231X surname: Tarancón-Álvarez fullname: Tarancón-Álvarez, Pedro email: pedro.tarancon@fqa.ub.edu organization: Departament de Física Quántica i Astrofísica, Universitat de Barcelona, Institut de Ciències del Cosmos (ICC), Universitat de Barcelona – sequence: 2 givenname: Pablo orcidid: 0009-0003-7249-9023 surname: Tejerina-Pérez fullname: Tejerina-Pérez, Pablo email: pablo.tejerina@icc.ub.edu organization: Departament de Física Quántica i Astrofísica, Universitat de Barcelona, Institut de Ciències del Cosmos (ICC), Universitat de Barcelona – sequence: 3 givenname: Raul orcidid: 0000-0002-3370-3103 surname: Jimenez fullname: Jimenez, Raul organization: Institut de Ciències del Cosmos (ICC), Universitat de Barcelona, Institució Catalana de Recerca i Estudis Avançats, ICREA – sequence: 4 givenname: Pavlos surname: Protopapas fullname: Protopapas, Pavlos organization: Institute for Applied Computational Science, Harvard University
BookMark	eNp9UU1LxTAQDKLg5x_wFPBc3SRtkx5F_Hgg6kHPIba7zzz6mmfSCvrrba2oJw_LbMLMbLKzz7a70CFjxwJOBShzlnIJUGQgp5K5ycQW25OqqjJVFrD9p99lRymtAECKHLQq99j9JZGvPXY9f1jc3SX-5h1fD23vsxd0DR86vw7N0LrIIy4n9B-u96HjgXj_gjyFdpjOiaeNq_GQ7ZBrEx594wF7urp8vLjJbu-vFxfnt1mtjOwzmVcFggE0iqRAoRvKFRUaZWPIgSYtTCMQx-9RqV2tpDJYSVOTQ01I6oAtZt8muJXdRL928d0G5-3XRYhL62Lv6xZtoUFXRGUjn2VeNlSRVg4qyGFcybMsR6-T2WsTw-uAqberMMRufL5V0-pUqaQeWXJm1TGkFJF-pgqwUw52zsGOOdivHKwYRWoWpZHcLTH-Wv-j-gSegIu2
Cites_doi	10.1109/ICDM58522.2023.00065 10.48550/arXiv.1908.02729 10.48550/arXiv.2311.14931 10.1007/JHEP08(2018)034 10.1103/PhysRevD.78.086007 10.1016/j.neunet.2024.106826 10.48550/arXiv.2010.15201 10.1038/s42254-021-00314-5 10.1016/j.jcp.2018.08.029 10.21105/joss.01931 10.48550/arXiv.2211.00214 10.1016/j.jcp.2018.10.045 10.48550/arXiv.2310.17053 10.1016/j.engappai.2023.107324 10.48550/arXiv.2111.08410 10.48550/arXiv.2301.02152 10.1109/TMAG.2023.3247023 10.48550/arXiv.2301.07641 10.1038/s42256-021-00302-5 10.1007/JHEP07(2024)087 10.1109/TNNLS.2019.2919705 10.1137/19M1274067 10.1016/j.jcp.2019.05.024 10.48550/arXiv.1906.01563 10.1002/cnm.1640100303 10.1155/2014/787412 10.48550/arXiv.2010.08895 10.1007/978-3-642-05221-7 10.48550/arXiv.2007.04216 10.48550/arXiv.2110.11286 10.1103/PhysRevE.101.062207 10.48550/arXiv.2407.10836 10.48550/arXiv.2501.17281 10.1103/PhysRevE.105.065305 10.1088/1361-6382/aca386 10.1016/j.cma.2024.117628 10.1090/S0002-9904-1955-09934-8 10.1109/72.870037 10.48550/arXiv.2006.14372
ContentType	Journal Article
Copyright	The Author(s) 2025 The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by/4.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Copyright_xml	– notice: The Author(s) 2025 – notice: The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by/4.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
DBID	C6C AAYXX CITATION 3V. 7XB 88I 8FE 8FG 8FK ABJCF ABUWG AFKRA AZQEC BENPR BGLVJ CCPQU DWQXO GNUQQ HCIFZ L6V M2P M7S PHGZM PHGZT PIMPY PKEHL PQEST PQGLB PQQKQ PQUKI PTHSS Q9U DOA
DOI	10.1038/s42005-025-02248-1
DatabaseName	Springer Nature OA Free Journals CrossRef ProQuest Central (Corporate) ProQuest Central (purchase pre-March 2016) Science Database (Alumni Edition) ProQuest SciTech Collection ProQuest Technology Collection ProQuest Central (Alumni) (purchase pre-March 2016) Materials Science & Engineering Collection ProQuest Central (Alumni) ProQuest Central UK/Ireland ProQuest Central Essentials ProQuest Central ProQuest Technology Collection ProQuest One ProQuest Central ProQuest Central Student SciTech Premium Collection ProQuest Engineering Collection Science Database (via ProQuest SciTech Premium Collection) Engineering Database ProQuest Central Premium ProQuest One Academic ProQuest Publicly Available Content Database ProQuest One Academic Middle East (New) ProQuest One Academic Eastern Edition (DO NOT USE) ProQuest One Applied & Life Sciences ProQuest One Academic (retired) ProQuest One Academic UKI Edition Engineering Collection ProQuest Central Basic DOAJ Directory of Open Access Journals
DatabaseTitle	CrossRef Publicly Available Content Database ProQuest Central Student Technology Collection ProQuest One Academic Middle East (New) ProQuest Central Essentials ProQuest Central (Alumni Edition) SciTech Premium Collection ProQuest One Community College ProQuest Central ProQuest One Applied & Life Sciences ProQuest Engineering Collection ProQuest Central Korea ProQuest Central (New) Engineering Collection Engineering Database ProQuest Science Journals (Alumni Edition) ProQuest Central Basic ProQuest Science Journals ProQuest One Academic Eastern Edition ProQuest Technology Collection ProQuest SciTech Collection ProQuest One Academic UKI Edition Materials Science & Engineering Collection ProQuest One Academic ProQuest One Academic (New) ProQuest Central (Alumni)
DatabaseTitleList	Publicly Available Content Database CrossRef
Database_xml	– sequence: 1 dbid: DOA name: DOAJ Directory of Open Access Journals url: https://www.doaj.org/ sourceTypes: Open Website – sequence: 2 dbid: PIMPY name: Publicly Available Content Database url: http://search.proquest.com/publiccontent sourceTypes: Aggregation Database
DeliveryMethod	fulltext_linktorsrc
Discipline	Physics
EISSN	2399-3650
EndPage	14
ExternalDocumentID	oai_doaj_org_article_57079ff6d2b246df9f73a09040365b26 10_1038_s42005_025_02248_1
GroupedDBID	0R~ 88I AAFWJ AAJSJ AARCD AASML ABDBF ABJCF ABUWG ACGFS ACUHS ADBBV ADMLS AFKRA AFPKN ALMA_UNASSIGNED_HOLDINGS AZQEC BCNDV BENPR BGLVJ C6C CCPQU DWQXO EBLON EBS GNUQQ GROUPED_DOAJ HCIFZ M2P M7S M~E NAO O9- OK1 PHGZM PHGZT PIMPY PQGLB PTHSS RNT SNYQT AAYXX AFFHD CITATION 3V. 7XB 8FE 8FG 8FK L6V PKEHL PQEST PQQKQ PQUKI PUEGO Q9U
ID	FETCH-LOGICAL-c382t-2495e080e83f21e17df43f57e2d8fa07f718d1ee038f67ac3238e928cfae7fef3
IEDL.DBID	DOA
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001550331800002&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	2399-3650
IngestDate	Mon Nov 10 04:32:34 EST 2025 Sat Aug 23 13:00:29 EDT 2025 Sat Nov 29 07:35:33 EST 2025 Sat Aug 16 01:10:34 EDT 2025
IsDoiOpenAccess	true
IsOpenAccess	true
IsPeerReviewed	true
IsScholarly	true
Issue	1
Language	English
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c382t-2495e080e83f21e17df43f57e2d8fa07f718d1ee038f67ac3238e928cfae7fef3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ORCID	0009-0004-5774-231X 0000-0002-3370-3103 0009-0003-7249-9023
OpenAccessLink	https://doaj.org/article/57079ff6d2b246df9f73a09040365b26
PQID	3239936327
PQPubID	4669724
PageCount	14
ParticipantIDs	doaj_primary_oai_doaj_org_article_57079ff6d2b246df9f73a09040365b26 proquest_journals_3239936327 crossref_primary_10_1038_s42005_025_02248_1 springer_journals_10_1038_s42005_025_02248_1
PublicationCentury	2000
PublicationDate	2025-08-15
PublicationDateYYYYMMDD	2025-08-15
PublicationDate_xml	– month: 08 year: 2025 text: 2025-08-15 day: 15
PublicationDecade	2020
PublicationPlace	London
PublicationPlace_xml	– name: London
PublicationTitle	Communications physics
PublicationTitleAbbrev	Commun Phys
PublicationYear	2025
Publisher	Nature Publishing Group UK Nature Publishing Group Nature Portfolio
Publisher_xml	– name: Nature Publishing Group UK – name: Nature Publishing Group – name: Nature Portfolio
References	2248_CR9 M Khalique (2248_CR2) 2014; 2014 J Sirignano (2248_CR31) 2018; 375 2248_CR4 2248_CR3 P Kaul (2248_CR14) 2020; 31 M Mattheakis (2248_CR30) 2022; 105 Y Bea (2248_CR39) 2024; 07 2248_CR26 GE Karniadakis (2248_CR7) 2021; 3 2248_CR25 L Lu (2248_CR8) 2021; 3 2248_CR22 2248_CR23 2248_CR21 I Lagaris (2248_CR29) 2000; 11 F Chen (2248_CR5) 2020; 5 JD Lambert (2248_CR12) 1991 A Choudhary (2248_CR33) 2019; 101 M Berardi (2248_CR24) 2025; 435 E Pinney (2248_CR1) 1955; 61 M Baldan (2248_CR27) 2023; 59 SS Gubser (2248_CR41) 2008; 78 2248_CR15 E Cardoso-Bihlo (2248_CR20) 2025; 181 2248_CR16 2248_CR38 2248_CR35 2248_CR36 2248_CR11 R Carballo-Rubio (2248_CR37) 2022; 39 M Raissi (2248_CR13) 2019; 378 2248_CR34 Y Bea (2248_CR40) 2018; 08 2248_CR10 Y Zhu (2248_CR32) 2019; 394 F Sahli Costabal (2248_CR19) 2024; 127 M Dissanayake (2248_CR28) 1994; 10 2248_CR17 2248_CR18 L Lu (2248_CR6) 2021; 63
References_xml	– ident: 2248_CR17 doi: 10.1109/ICDM58522.2023.00065 – ident: 2248_CR38 doi: 10.48550/arXiv.1908.02729 – ident: 2248_CR25 doi: 10.48550/arXiv.2311.14931 – volume: 08 start-page: 034 year: 2018 ident: 2248_CR40 publication-title: JHEP doi: 10.1007/JHEP08(2018)034 – volume: 78 start-page: 086007 year: 2008 ident: 2248_CR41 publication-title: Phys. Rev. D. doi: 10.1103/PhysRevD.78.086007 – volume: 181 start-page: 106826 year: 2025 ident: 2248_CR20 publication-title: Neural Netw. doi: 10.1016/j.neunet.2024.106826 – ident: 2248_CR34 doi: 10.48550/arXiv.2010.15201 – volume: 3 start-page: 422 year: 2021 ident: 2248_CR7 publication-title: Nat. Rev. Phys. doi: 10.1038/s42254-021-00314-5 – volume: 375 start-page: 1339 year: 2018 ident: 2248_CR31 publication-title: J. Comput. Phys. doi: 10.1016/j.jcp.2018.08.029 – volume: 5 start-page: 1931 year: 2020 ident: 2248_CR5 publication-title: J. Open Source Softw. doi: 10.21105/joss.01931 – ident: 2248_CR26 doi: 10.48550/arXiv.2211.00214 – volume-title: Numerical Methods for Ordinary Differential Systems: The Initial Value Problem year: 1991 ident: 2248_CR12 – volume: 378 start-page: 686 year: 2019 ident: 2248_CR13 publication-title: J. Comput. Phys. doi: 10.1016/j.jcp.2018.10.045 – ident: 2248_CR18 doi: 10.48550/arXiv.2310.17053 – volume: 127 start-page: 107324 year: 2024 ident: 2248_CR19 publication-title: Eng. Appl. Artif. Intell. doi: 10.1016/j.engappai.2023.107324 – ident: 2248_CR15 doi: 10.48550/arXiv.2111.08410 – ident: 2248_CR21 doi: 10.48550/arXiv.2301.02152 – volume: 59 start-page: 1 year: 2023 ident: 2248_CR27 publication-title: IEEE Trans. Magn. doi: 10.1109/TMAG.2023.3247023 – ident: 2248_CR36 doi: 10.48550/arXiv.2301.07641 – volume: 3 start-page: 218 year: 2021 ident: 2248_CR8 publication-title: Nat. Mach. Intell. doi: 10.1038/s42256-021-00302-5 – volume: 07 start-page: 087 year: 2024 ident: 2248_CR39 publication-title: JHEP doi: 10.1007/JHEP07(2024)087 – volume: 31 start-page: 1410 year: 2020 ident: 2248_CR14 publication-title: IEEE Trans. Neural Netw. Learn. Syst. doi: 10.1109/TNNLS.2019.2919705 – volume: 63 start-page: 208 year: 2021 ident: 2248_CR6 publication-title: SIAM Rev. doi: 10.1137/19M1274067 – volume: 394 start-page: 56 year: 2019 ident: 2248_CR32 publication-title: J. Comput. Phys. doi: 10.1016/j.jcp.2019.05.024 – ident: 2248_CR35 doi: 10.48550/arXiv.1906.01563 – volume: 10 start-page: 195 year: 1994 ident: 2248_CR28 publication-title: Commun. Numer. Methods Eng. doi: 10.1002/cnm.1640100303 – volume: 2014 start-page: 1 year: 2014 ident: 2248_CR2 publication-title: Math. Probl. Eng. doi: 10.1155/2014/787412 – ident: 2248_CR9 doi: 10.48550/arXiv.2010.08895 – ident: 2248_CR3 – ident: 2248_CR10 doi: 10.1007/978-3-642-05221-7 – ident: 2248_CR16 doi: 10.48550/arXiv.2007.04216 – ident: 2248_CR22 doi: 10.48550/arXiv.2110.11286 – volume: 101 start-page: 062207 year: 2019 ident: 2248_CR33 publication-title: Phys. Rev. E doi: 10.1103/PhysRevE.101.062207 – ident: 2248_CR23 doi: 10.48550/arXiv.2407.10836 – ident: 2248_CR11 doi: 10.48550/arXiv.2501.17281 – volume: 105 start-page: 065305 year: 2022 ident: 2248_CR30 publication-title: Phys. Rev. E doi: 10.1103/PhysRevE.105.065305 – volume: 39 start-page: 243001 year: 2022 ident: 2248_CR37 publication-title: Class. Quant. Grav. doi: 10.1088/1361-6382/aca386 – volume: 435 start-page: 117628 year: 2025 ident: 2248_CR24 publication-title: Comput. Methods Appl. Mech. Eng. doi: 10.1016/j.cma.2024.117628 – volume: 61 start-page: 373 year: 1955 ident: 2248_CR1 publication-title: Bull. N. Ser. Am. Math. Soc. doi: 10.1090/S0002-9904-1955-09934-8 – volume: 11 start-page: 1041 year: 2000 ident: 2248_CR29 publication-title: IEEE Trans. Neural Netw. doi: 10.1109/72.870037 – ident: 2248_CR4 doi: 10.48550/arXiv.2006.14372
SSID	ssj0002140736
Score	2.307372
Snippet	Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method to tackle... Abstract Non-linear differential equations are a fundamental tool to describe different phenomena in nature. However, we still lack a well-established method...
SourceID	doaj proquest crossref springer
SourceType	Open Website Aggregation Database Index Database Publisher
StartPage	335
SubjectTerms	639/705/1042 639/766/259 Accuracy Algorithms Boundary conditions Einstein equations Geometry Inverse problems Linear equations Machine learning Neural networks Nonlinear differential equations Numerical analysis Ordinary differential equations Physics Physics and Astronomy Regularization
SummonAdditionalLinks	– databaseName: Science Database (via ProQuest SciTech Premium Collection) dbid: M2P link: http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV07T8MwELaggMTCG1Fe8sAGURPbiZ0JAQIxQGEAic1ybB9iaSEp_f2c3bQVSLCwJpbl5Dvfw3e-j5ATWxYK0qpIKiUqDFBMmhjlIBEWUE-WDgOhyFpyJ_t99fJSPrYHbk1bVjnViVFRu6ENZ-Q9Hi5h8oIzef7-kQTWqJBdbSk0FskSejZZKOm6Z4-zMxaG0YMM2cn9ye1y1WtEbL0ZOFyD8cIA6ps9im37v_maP9Kj0ercrP93vRtkrfU36cVEQDbJgh9skZVY92mbbfJwHVtIoOWhGN_3Gzp-MzQWGSaopR39HCCWLpSq0jqy1tftvU06BIq-I52JLkXVZP0Oeb65frq6TVqOhcRyxUZJoJ726DV6xYFlPpMOBIdceuYUmFQC2i6XeY8_DwppLH6W8iVTFoyX4IHvks5gOPB7hGalgdQ5IQCMwAkrFALPKwxIrPC5zLrkdPqn9fuklYaOKXCu9AQXjbjoiIvG0ZcBjNnI0AY7PhjWr7rdVToP_f0ACscqJgoHJUhu0hIVEy_yihVdcjgFR7d7s9FzZLrkbArv_PXvS9r_e7YDssqiYKFU5YekM6o__RFZtuPRW1MfR8n8Ar9g6g4 priority: 102 providerName: ProQuest
Title	Efficient PINNs via multi-head unimodular regularization of the solutions space
URI	https://link.springer.com/article/10.1038/s42005-025-02248-1 https://www.proquest.com/docview/3239936327 https://doaj.org/article/57079ff6d2b246df9f73a09040365b26
Volume	8
WOSCitedRecordID	wos001550331800002&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: 2399-3650 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002140736 issn: 2399-3650 databaseCode: DOA dateStart: 20180101 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: 2399-3650 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002140736 issn: 2399-3650 databaseCode: M~E dateStart: 20180101 isFulltext: true titleUrlDefault: https://road.issn.org providerName: ISSN International Centre – providerCode: PRVPQU databaseName: Engineering Database customDbUrl: eissn: 2399-3650 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002140736 issn: 2399-3650 databaseCode: M7S dateStart: 20250101 isFulltext: true titleUrlDefault: http://search.proquest.com providerName: ProQuest – providerCode: PRVPQU databaseName: ProQuest Central customDbUrl: eissn: 2399-3650 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002140736 issn: 2399-3650 databaseCode: BENPR dateStart: 20250101 isFulltext: true titleUrlDefault: https://www.proquest.com/central providerName: ProQuest – providerCode: PRVPQU databaseName: Publicly Available Content Database customDbUrl: eissn: 2399-3650 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002140736 issn: 2399-3650 databaseCode: PIMPY dateStart: 20250101 isFulltext: true titleUrlDefault: http://search.proquest.com/publiccontent providerName: ProQuest – providerCode: PRVPQU databaseName: Science Database customDbUrl: eissn: 2399-3650 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002140736 issn: 2399-3650 databaseCode: M2P dateStart: 20250101 isFulltext: true titleUrlDefault: https://search.proquest.com/sciencejournals providerName: ProQuest
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV09T8MwELUQMLAgECAKBXlgg4jEdmJnpKgIJCgRH1KZLCf2SV1a1LT9_ZydFCgSYmHxEFuR8-5yd08-3xFyVuWZgrjMolKJEgmKiSOjLESiArSTuUUiFLqW3MvBQA2HefGt1ZfPCWvKAzfAIWGPZQ6QWVYykVnIQXIT56h7PEtLFopt44pvZMrbYIa8QfpzyaPmXrm6rEUouum7t3q3hdRpxROFgv0rUeaPg9Hgb252yHYbKNKrZoO7ZM2N98hjP1R8QEdBkY4ParoYGRpyAiM0qpbOxwi99ZmldBqazE_ba5Z0AhRDPfqpaRQtSeX2yetN_-X6NmpbIkQVV2wW-U7RDoM8pziwxCXSguCQSsesAhNLQFdjE-fwiyGTpuLokV3OVAXGSXDAD8j6eDJ2h4QmuYHYWiEAjMAXligzx0vkD5VwqUw65HwJj35vKl_ocGLNlW7A1AimDmBqXN3zCH6u9FWrwwOUpW5lqf-SZYd0l_jr9leqNfe3b3nGmeyQi6VMvqZ_39LRf2zpmGyxoDOoMGmXrM-mc3dCNqvFbFRPT8lGrz8onk6D8uH4wAo_ymecKe4eircPfpLdNQ
linkProvider	Directory of Open Access Journals
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMw1V1Lb9QwEB6VAoILb8S2BXyAE0RNbCd2DhVqoVVXXcIeitSbSWxP1ctu2WyL-FP8RsZOslWR4NYDp0hJ5Dz8-ZsZzwvgjS0LjWlTJI2WDRkodZrU2mEiLRJPlo4Modi1ZKKqSp-clNM1-DXkwoSwyoETI1G7uQ175NsiJGGKQnD14fx7ErpGBe_q0EKjg8WR__mDTLZ2Z_yJ5vct5wf7xx8Pk76rQGKF5sskNFv2pCd5LZBnPlMOpcBcee401qlCYmuXeZ8KjYWqLT1Y-5Jri7VX6FHQuLfgtgyVxUKoIJ-u9nQ4WSsqeEM3umx2vd3KWOoz9IwNwpIMtmvyL7YJuKbb_uGOjVLu4OH_9n8ewYNen2a73QJ4DGt-9gTuxrhW2z6FL_uxRAZJVjYdV1XLLs9qFoMoE5JCjl3MCKsuhOKyhT8Nxz4vlc2RkW7MVkuTEfVa_wy-3sjXPIf12XzmXwDLyhpT56RErCUN2BDIvWjI4LLS5yobwbthZs15VyrERBe_0KbDgSEcmIgDQ3fvhclf3RnKfMcT88Wp6VnD5KF-IWLheMNl4bBEJeq0JOIVRd7wYgRbAxhMzz2tuULCCN4PcLq6_PdX2vj3aK_h3uHx54mZjKujTbjPI6gJ0fkWrC8XF_4l3LGXy7N28SquCgbfbhpmvwG9V0gu
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=Efficient+PINNs+via+multi-head+unimodular+regularization+of+the+solutions+space&rft.jtitle=Communications+physics&rft.au=Pedro+Taranc%C3%B3n-%C3%81lvarez&rft.au=Pablo+Tejerina-P%C3%A9rez&rft.au=Raul+Jimenez&rft.au=Pavlos+Protopapas&rft.date=2025-08-15&rft.pub=Nature+Portfolio&rft.eissn=2399-3650&rft.volume=8&rft.issue=1&rft.spage=1&rft.epage=14&rft_id=info:doi/10.1038%2Fs42005-025-02248-1&rft.externalDBID=DOA&rft.externalDocID=oai_doaj_org_article_57079ff6d2b246df9f73a09040365b26
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2399-3650&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2399-3650&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2399-3650&client=summon