Respecting causality for training physics-informed neural networks

While the popularity of physics-informed neural networks (PINNs) is steadily rising, to this date PINNs have not been successful in simulating dynamical systems whose solution exhibits multi-scale, chaotic or turbulent behavior. In this work we attribute this shortcoming to the inability of existing...

Full description

Saved in:
Bibliographic Details
Published in:Computer methods in applied mechanics and engineering Vol. 421; no. C; p. 116813
Main Authors: Wang, Sifan, Sankaran, Shyam, Perdikaris, Paris
Format: Journal Article
Language:English
Published: Netherlands Elsevier B.V 01.03.2024
Elsevier
Subjects:
ISSN:0045-7825, 1879-2138
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract While the popularity of physics-informed neural networks (PINNs) is steadily rising, to this date PINNs have not been successful in simulating dynamical systems whose solution exhibits multi-scale, chaotic or turbulent behavior. In this work we attribute this shortcoming to the inability of existing PINNs formulations to respect the spatio-temporal causal structure that is inherent to the evolution of physical systems. We argue that this is a fundamental limitation and a key source of error that can ultimately steer PINN models to converge towards erroneous solutions. We address this pathology by proposing a simple re-formulation of PINNs loss functions that can explicitly account for physical causality during model training. We demonstrate that this simple modification alone is enough to introduce significant accuracy improvements, as well as a practical quantitative mechanism for assessing the convergence of a PINNs model. We provide state-of-the-art numerical results across a series of benchmarks for which existing PINNs formulations fail, including the chaotic Lorenz system, the Kuramoto–Sivashinsky equation in the chaotic regime, and the Navier–Stokes equations. To the best of our knowledge, this is the first time that PINNs have been successful in simulating such systems, introducing new opportunities for their applicability to problems of industrial complexity.
AbstractList While the popularity of physics-informed neural networks (PINNs) is steadily rising, to this date PINNs have not been successful in simulating dynamical systems whose solution exhibits multi-scale, chaotic or turbulent behavior. In this work we attribute this shortcoming to the inability of existing PINNs formulations to respect the spatio-temporal causal structure that is inherent to the evolution of physical systems. We argue that this is a fundamental limitation and a key source of error that can ultimately steer PINN models to converge towards erroneous solutions. We address this pathology by proposing a simple re-formulation of PINNs loss functions that can explicitly account for physical causality during model training. We demonstrate that this simple modification alone is enough to introduce significant accuracy improvements, as well as a practical quantitative mechanism for assessing the convergence of a PINNs model. We provide state-of-the-art numerical results across a series of benchmarks for which existing PINNs formulations fail, including the chaotic Lorenz system, the Kuramoto–Sivashinsky equation in the chaotic regime, and the Navier–Stokes equations. To the best of our knowledge, this is the first time that PINNs have been successful in simulating such systems, introducing new opportunities for their applicability to problems of industrial complexity.
ArticleNumber 116813
Author Perdikaris, Paris
Sankaran, Shyam
Wang, Sifan
Author_xml – sequence: 1
  givenname: Sifan
  orcidid: 0000-0002-7721-6884
  surname: Wang
  fullname: Wang, Sifan
  email: sifanw@sas.upenn.edu
  organization: Graduate Group in Applied Mathematics and Computational Science, University of Pennsylvania, Philadelphia, PA 19104, United States of America
– sequence: 2
  givenname: Shyam
  orcidid: 0000-0002-1583-2147
  surname: Sankaran
  fullname: Sankaran, Shyam
  email: shyamss@seas.upenn.edu
  organization: Department of Mechanical Engineering and Applied Mechanics, University of Pennsylvania, Philadelphia, PA 19104, United States of America
– sequence: 3
  givenname: Paris
  orcidid: 0000-0002-2816-3229
  surname: Perdikaris
  fullname: Perdikaris, Paris
  email: pgp@seas.upenn.edu
  organization: Department of Mechanical Engineering and Applied Mechanics, University of Pennsylvania, Philadelphia, PA 19104, United States of America
BackLink https://www.osti.gov/biblio/2290430$$D View this record in Osti.gov
BookMark eNp9kM1KAzEUhYNUsK0-gLvB_Yz5m06CKy1WhYIgug6ZTGJTp0lJUqVvb4Zx5aJ3c-Fwvsu5ZwYmzjsNwDWCFYJocbut1E5WGGJaIbRgiJyBKWINLzEibAKmENK6bBiuL8Asxi3MwxCegoc3HfdaJes-CyUPUfY2HQvjQ5GCtG6Q95tjtCqW1mV5p7vC6UOQfV7px4eveAnOjeyjvvrbc_CxenxfPpfr16eX5f26VJSQVNYdRQzJTknJYdfwNqfTmOTEBjKjIe0IbZWhhNUtR3WWedu0nBpWk5qblszBzXjXx2RFVDZptVHeuRxfYMwhJTCbmtGkgo8xaCOyTybr3fBPLxAUQ19iK3JfYuhLjH1lEv0j98HuZDieZO5GRue_v60OQyztlO5sGFJ13p6gfwFPMIUJ
CitedBy_id crossref_primary_10_1007_s10915_024_02709_9
crossref_primary_10_1063_5_0273148
crossref_primary_10_1016_j_compbiomed_2024_109476
crossref_primary_10_1002_nme_7637
crossref_primary_10_1016_j_cma_2024_117586
crossref_primary_10_1098_rspa_2024_0270
crossref_primary_10_1088_1674_1056_adacd0
crossref_primary_10_3390_e26050396
crossref_primary_10_1016_j_matdes_2025_114284
crossref_primary_10_1016_j_engappai_2024_108156
crossref_primary_10_1016_j_cpc_2024_109422
crossref_primary_10_1016_j_neunet_2025_107983
crossref_primary_10_1016_j_apm_2025_116232
crossref_primary_10_1016_j_aei_2025_103581
crossref_primary_10_3390_math13071057
crossref_primary_10_1063_5_0252852
crossref_primary_10_1016_j_cma_2024_117691
crossref_primary_10_1063_5_0228104
crossref_primary_10_1017_jfm_2025_10448
crossref_primary_10_1016_j_buildenv_2025_113616
crossref_primary_10_1016_j_jhydrol_2024_131261
crossref_primary_10_1038_s41598_024_67597_3
crossref_primary_10_1109_ACCESS_2024_3504962
crossref_primary_10_1016_j_jcmds_2024_100107
crossref_primary_10_1016_j_addma_2025_104881
crossref_primary_10_1016_j_jcp_2025_114226
crossref_primary_10_1016_j_apenergy_2025_126040
crossref_primary_10_1088_2632_2153_ad5a60
crossref_primary_10_1016_j_commatsci_2024_113502
crossref_primary_10_1063_5_0220392
crossref_primary_10_1016_j_neunet_2024_106998
crossref_primary_10_1109_TEMC_2024_3490699
crossref_primary_10_1038_s44384_025_00021_w
crossref_primary_10_1088_2632_2153_ad450f
crossref_primary_10_1007_s11071_025_11158_4
crossref_primary_10_1016_j_compgeo_2025_107055
crossref_primary_10_1080_14697688_2025_2477673
crossref_primary_10_1016_j_cma_2025_117956
crossref_primary_10_1002_bit_28851
crossref_primary_10_1016_j_compind_2025_104304
crossref_primary_10_1016_j_enganabound_2025_106200
crossref_primary_10_1016_j_cma_2025_117851
crossref_primary_10_1016_j_cam_2025_116858
crossref_primary_10_1016_j_engappai_2024_109886
crossref_primary_10_1007_s00466_023_02435_3
crossref_primary_10_1007_s00521_025_11554_2
crossref_primary_10_1007_s11071_025_11577_3
crossref_primary_10_1016_j_cma_2025_118260
crossref_primary_10_1109_TGRS_2025_3581638
crossref_primary_10_1016_j_camwa_2025_09_014
crossref_primary_10_1016_j_enganabound_2025_106207
crossref_primary_10_1016_j_petsci_2025_07_008
crossref_primary_10_1137_23M1556320
crossref_primary_10_1371_journal_pone_0315762
crossref_primary_10_1016_j_cma_2025_118356
crossref_primary_10_1016_j_matcom_2024_10_039
crossref_primary_10_1088_1572_9494_adcc8e
crossref_primary_10_1115_1_4068396
crossref_primary_10_1016_j_finel_2024_104305
crossref_primary_10_3390_app15168863
crossref_primary_10_1016_j_physd_2025_134878
crossref_primary_10_1126_sciadv_ads5236
crossref_primary_10_1016_j_jcp_2025_114156
crossref_primary_10_1016_j_neucom_2025_131589
crossref_primary_10_1016_j_engappai_2025_110547
crossref_primary_10_1016_j_jcp_2025_113860
crossref_primary_10_1063_5_0285282
crossref_primary_10_1016_j_geoderma_2024_117094
crossref_primary_10_1007_s44379_025_00015_1
crossref_primary_10_1016_j_ijmecsci_2025_110374
crossref_primary_10_1088_1402_4896_adeaf8
crossref_primary_10_1063_5_0211398
crossref_primary_10_1080_15502287_2024_2440420
crossref_primary_10_1063_5_0279565
crossref_primary_10_1103_ytyy_pvys
crossref_primary_10_1063_5_0276518
crossref_primary_10_1016_j_cma_2025_117764
crossref_primary_10_1177_10812865251362165
crossref_primary_10_3390_fluids10070184
crossref_primary_10_1016_j_ocemod_2025_102601
crossref_primary_10_3390_math13091515
crossref_primary_10_1016_j_engappai_2025_111098
crossref_primary_10_1016_j_jcp_2025_114370
crossref_primary_10_1016_j_physa_2025_130759
crossref_primary_10_1016_j_neunet_2025_107841
crossref_primary_10_1016_j_watres_2025_123427
crossref_primary_10_1016_j_cpc_2025_109672
crossref_primary_10_1080_17486025_2025_2502029
crossref_primary_10_1016_j_brainresbull_2025_111318
crossref_primary_10_1007_s44379_025_00016_0
crossref_primary_10_1016_j_jocs_2025_102577
crossref_primary_10_1063_5_0287489
crossref_primary_10_1007_s42791_025_00101_8
crossref_primary_10_1134_S106456242460194X
crossref_primary_10_1007_s00366_025_02174_4
crossref_primary_10_1016_j_enganabound_2025_106363
crossref_primary_10_1016_j_jcp_2025_113837
crossref_primary_10_1109_TII_2024_3507213
crossref_primary_10_3390_math13111882
crossref_primary_10_1016_j_jcp_2024_113656
crossref_primary_10_1007_s00033_025_02515_9
crossref_primary_10_1016_j_jcp_2025_113799
crossref_primary_10_1016_j_matcom_2024_10_043
Cites_doi 10.1016/j.cma.2019.112623
10.1016/j.neucom.2021.10.036
10.1137/20M1318043
10.1016/j.array.2021.100110
10.1038/s41586-020-2649-2
10.1109/JPROC.2021.3058954
10.1126/science.aaw4741
10.3389/fphy.2020.00042
10.1016/j.jcp.2021.110768
10.1016/j.jcp.2021.110242
10.1016/0094-5765(77)90096-0
10.1145/3241036
10.1175/1520-0469(1963)020<0130:DNF>2.0.CO;2
10.1364/OE.384875
10.4208/cicp.OA-2020-0164
10.1016/j.jcp.2019.109136
10.1111/mice.12685
10.1002/aic.690381003
10.1143/PTP.55.356
10.1103/PhysRevE.104.025205
10.1109/72.712178
10.1103/PhysRevFluids.4.124501
10.1016/j.jcp.2020.109914
10.1016/j.jcp.2018.10.045
10.1016/j.cma.2021.113938
10.1016/j.cma.2021.114474
10.1007/s10921-020-00705-1
10.1371/journal.pcbi.1007575
10.1073/pnas.2101784118
ContentType Journal Article
Copyright 2024 Elsevier B.V.
Copyright_xml – notice: 2024 Elsevier B.V.
DBID AAYXX
CITATION
OTOTI
DOI 10.1016/j.cma.2024.116813
DatabaseName CrossRef
OSTI.GOV
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Applied Sciences
Engineering
EISSN 1879-2138
ExternalDocumentID 2290430
10_1016_j_cma_2024_116813
S0045782524000690
GroupedDBID --K
--M
-~X
.DC
.~1
0R~
1B1
1~.
1~5
4.4
457
4G.
5GY
5VS
7-5
71M
8P~
9JN
AABNK
AACTN
AAEDT
AAEDW
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAXKI
AAXUO
AAYFN
ABAOU
ABBOA
ABFNM
ABJNI
ABMAC
ACDAQ
ACGFS
ACIWK
ACRLP
ACZNC
ADBBV
ADEZE
ADGUI
ADTZH
AEBSH
AECPX
AEKER
AENEX
AFJKZ
AFKWA
AFTJW
AGHFR
AGUBO
AGYEJ
AHHHB
AHJVU
AHZHX
AIALX
AIEXJ
AIGVJ
AIKHN
AITUG
AJOXV
AKRWK
ALMA_UNASSIGNED_HOLDINGS
AMFUW
AMRAJ
AOUOD
ARUGR
AXJTR
BJAXD
BKOJK
BLXMC
CS3
DU5
EBS
EFJIC
EO8
EO9
EP2
EP3
F5P
FDB
FIRID
FNPLU
FYGXN
G-Q
GBLVA
GBOLZ
IHE
J1W
JJJVA
KOM
LG9
LY7
M41
MHUIS
MO0
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
PQQKQ
Q38
RIG
RNS
ROL
RPZ
SDF
SDG
SDP
SES
SEW
SPC
SPCBC
SST
SSV
SSW
SSZ
T5K
TN5
WH7
XPP
ZMT
~02
~G-
29F
9DU
AAQXK
AATTM
AAYWO
AAYXX
ABEFU
ABWVN
ABXDB
ACLOT
ACNNM
ACRPL
ACVFH
ADCNI
ADIYS
ADJOM
ADMUD
ADNMO
AEIPS
AEUPX
AFPUW
AGQPQ
AI.
AIGII
AIIUN
AKBMS
AKYEP
ANKPU
APXCP
ASPBG
AVWKF
AZFZN
CITATION
EFKBS
EFLBG
EJD
FEDTE
FGOYB
G-2
HLZ
HVGLF
HZ~
R2-
SBC
SET
VH1
VOH
WUQ
ZY4
~HD
AAIAV
ABYKQ
ACAZW
OTOTI
ID FETCH-LOGICAL-c433t-5d4181adcaa90d79b879e23024f08fe04d34bcf4385b91524f9b7b94f85359fb3
ISICitedReferencesCount 101
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001179213600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0045-7825
IngestDate Mon Feb 12 04:55:34 EST 2024
Tue Nov 18 22:25:03 EST 2025
Sat Nov 29 06:16:58 EST 2025
Sat Oct 26 15:42:17 EDT 2024
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue C
Keywords Deep learning
Computational physics
Partial differential equations
Chaotic systems
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c433t-5d4181adcaa90d79b879e23024f08fe04d34bcf4385b91524f9b7b94f85359fb3
Notes USDOE Advanced Research Projects Agency - Energy (ARPA-E)
ORCID 0000-0002-2816-3229
0000-0002-7721-6884
0000-0002-1583-2147
0000000228163229
0000000277216884
0000000215832147
OpenAccessLink https://www.osti.gov/biblio/2339526
ParticipantIDs osti_scitechconnect_2290430
crossref_citationtrail_10_1016_j_cma_2024_116813
crossref_primary_10_1016_j_cma_2024_116813
elsevier_sciencedirect_doi_10_1016_j_cma_2024_116813
PublicationCentury 2000
PublicationDate 2024-03-01
2024-03-00
PublicationDateYYYYMMDD 2024-03-01
PublicationDate_xml – month: 03
  year: 2024
  text: 2024-03-01
  day: 01
PublicationDecade 2020
PublicationPlace Netherlands
PublicationPlace_xml – name: Netherlands
PublicationTitle Computer methods in applied mechanics and engineering
PublicationYear 2024
Publisher Elsevier B.V
Elsevier
Publisher_xml – name: Elsevier B.V
– name: Elsevier
References Shukla, Jagtap, Karniadakis (b61) 2021
Raissi, Perdikaris, Karniadakis (b11) 2019; 378
Kaddour, Lynch, Liu, Kusner, Silva (b44) 2022
Wang, Wang, Perdikaris (b9) 2021; 384
Bu, Karpatne (b18) 2021
Wang, Perdikaris (b39) 2021
Li, Zheng, Kovachki, Jin, Chen, Liu, Azizzadenesheli, Anandkumar (b41) 2021
Harris, Millman, van der Walt, Gommers, Virtanen, Cournapeau, Wieser, Taylor, Berg, Smith (b64) 2020; 585
Hennigh, Narasimhan, Nabian, Subramaniam, Tangsali, Fang, Rietmann, Byeon, Choudhry (b37) 2021
Liang, Lyu, Wang, Yang (b20) 2021
Jung, Tian, Bareinboim (b45) 2020; vol. 33
Byrd, Lipton (b47) 2019
McClenny, Braga-Neto (b14) 2020
Krishnapriyan, Gholami, Zhe, Kirby, Mahoney (b24) 2021
Lagaris, Likas, Fotiadis (b28) 1998; 9
Sivashinsky (b58) 1977; 4
Pearl (b42) 2019; 62
Karniadakis, Kevrekidis, Lu, Perdikaris, Wang, Yang (b10) 2021
Kuramoto, Tsuzuki (b57) 1976; 55
Jagtap, Kawaguchi, Karniadakis (b23) 2020; 404
Jagtap, Shin, Kawaguchi, Karniadakis (b19) 2022; 468
Mattey, Ghosh (b25) 2022; 390
Griewank, Walther (b31) 2008
Dong, Ni (b50) 2021; 435
Wight, Zhao (b16) 2020
Wang, Wang, Perdikaris (b40) 2021
Strauss (b26) 2007
Hunter (b63) 2007; 9
Wang, Yu, Perdikaris (b13) 2022; 449
Wang, Wang, Perdikaris (b38) 2021
Du, Zaki (b48) 2021
Evans, American Mathematical Society (b27) 1998
Shukla, Di Leoni, Blackshire, Sparkman, Karniadakis (b6) 2020; 39
Kharazmi, Zhang, Karniadakis (b30) 2019
Bettencourt, Johnson, Duvenaud (b53) 2019
Schaarschmidt, Grewe, Vytiniotis, Paszke, Schmid, Norman, Molloy, Godwin, Rink, Nair (b60) 2021
Raissi, Yazdani, Karniadakis (b1) 2020; 367
Jagtap, Karniadakis (b21) 2020; 28
Kingma, Ba (b32) 2014
Vadyala, Betgeri, Betgeri (b49) 2022; 13
Nabian, Gladstone, Meidani (b17) 2021
Li, Beirami, Sanjabi, Smith (b46) 2020
Yazdani, Lu, Raissi, Karniadakis (b4) 2020; 16
Lorenz (b56) 1963; 20
Moseley, Markham, Nissen-Meyer (b22) 2021
Xavier Glorot, Yoshua Bengio, Understanding the difficulty of training deep feedforward neural networks, in: Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 2010, pp. 249–256.
Raissi, Babaee, Givi (b29) 2019; 4
Bradbury, Frostig, Hawkins, Johnson, Leary, Maclaurin, Necula, Paszke, VanderPlas, Wanderman-Milne, Zhang (b35) 2018
Mathews, Francisquez, Hughes, Hatch, Zhu, Rogers (b2) 2021; 104
Psichogios, Ungar (b54) 1992; 38
Kissas, Yang, Hwuang, Witschey, Detre, Perdikaris (b3) 2020; 358
Jacot, Gabriel, Hongler (b34) 2018
Sukumar, Srivastava (b51) 2021
Chen, Lu, Karniadakis, Dal Negro (b7) 2020; 28
Raissi (b59) 2018; 19
Lu, Meng, Mao, Karniadakis (b36) 2019
Maddu, Sturm, Müller, Sbalzarini (b15) 2021
Iserles (b33) 2009
Schölkopf, Locatello, Bauer, Ke, Kalchbrenner, Goyal, Bengio (b43) 2021; 109
Wang, Perdikaris (b5) 2021; 428
Wang, Teng, Perdikaris (b12) 2021; 43
Sahli Costabal, Yang, Perdikaris, Hurtado, Kuhl (b8) 2020; 8
Lu, Pestourie, Yao, Wang, Verdugo, Johnson (b52) 2021
Kochkov, Smith, Alieva, Wang, Brenner, Hoyer (b62) 2021; 118
Shukla (10.1016/j.cma.2024.116813_b6) 2020; 39
Dong (10.1016/j.cma.2024.116813_b50) 2021; 435
Kochkov (10.1016/j.cma.2024.116813_b62) 2021; 118
Mathews (10.1016/j.cma.2024.116813_b2) 2021; 104
Hunter (10.1016/j.cma.2024.116813_b63) 2007; 9
Yazdani (10.1016/j.cma.2024.116813_b4) 2020; 16
Wang (10.1016/j.cma.2024.116813_b9) 2021; 384
Du (10.1016/j.cma.2024.116813_b48) 2021
Sukumar (10.1016/j.cma.2024.116813_b51) 2021
Bu (10.1016/j.cma.2024.116813_b18) 2021
Kaddour (10.1016/j.cma.2024.116813_b44) 2022
Sahli Costabal (10.1016/j.cma.2024.116813_b8) 2020; 8
McClenny (10.1016/j.cma.2024.116813_b14) 2020
Iserles (10.1016/j.cma.2024.116813_b33) 2009
Kingma (10.1016/j.cma.2024.116813_b32) 2014
Jung (10.1016/j.cma.2024.116813_b45) 2020; vol. 33
Maddu (10.1016/j.cma.2024.116813_b15) 2021
Li (10.1016/j.cma.2024.116813_b41) 2021
Nabian (10.1016/j.cma.2024.116813_b17) 2021
Psichogios (10.1016/j.cma.2024.116813_b54) 1992; 38
Griewank (10.1016/j.cma.2024.116813_b31) 2008
Lu (10.1016/j.cma.2024.116813_b52) 2021
Shukla (10.1016/j.cma.2024.116813_b61) 2021
Krishnapriyan (10.1016/j.cma.2024.116813_b24) 2021
Jagtap (10.1016/j.cma.2024.116813_b19) 2022; 468
Chen (10.1016/j.cma.2024.116813_b7) 2020; 28
Raissi (10.1016/j.cma.2024.116813_b59) 2018; 19
Jacot (10.1016/j.cma.2024.116813_b34) 2018
Schölkopf (10.1016/j.cma.2024.116813_b43) 2021; 109
Bettencourt (10.1016/j.cma.2024.116813_b53) 2019
Kissas (10.1016/j.cma.2024.116813_b3) 2020; 358
Strauss (10.1016/j.cma.2024.116813_b26) 2007
Moseley (10.1016/j.cma.2024.116813_b22) 2021
Evans (10.1016/j.cma.2024.116813_b27) 1998
Raissi (10.1016/j.cma.2024.116813_b29) 2019; 4
Hennigh (10.1016/j.cma.2024.116813_b37) 2021
Wang (10.1016/j.cma.2024.116813_b39) 2021
10.1016/j.cma.2024.116813_b55
Wang (10.1016/j.cma.2024.116813_b5) 2021; 428
Wang (10.1016/j.cma.2024.116813_b13) 2022; 449
Kharazmi (10.1016/j.cma.2024.116813_b30) 2019
Raissi (10.1016/j.cma.2024.116813_b1) 2020; 367
Harris (10.1016/j.cma.2024.116813_b64) 2020; 585
Lu (10.1016/j.cma.2024.116813_b36) 2019
Li (10.1016/j.cma.2024.116813_b46) 2020
Kuramoto (10.1016/j.cma.2024.116813_b57) 1976; 55
Raissi (10.1016/j.cma.2024.116813_b11) 2019; 378
Byrd (10.1016/j.cma.2024.116813_b47) 2019
Bradbury (10.1016/j.cma.2024.116813_b35) 2018
Sivashinsky (10.1016/j.cma.2024.116813_b58) 1977; 4
Wang (10.1016/j.cma.2024.116813_b12) 2021; 43
Karniadakis (10.1016/j.cma.2024.116813_b10) 2021
Wang (10.1016/j.cma.2024.116813_b38) 2021
Jagtap (10.1016/j.cma.2024.116813_b23) 2020; 404
Pearl (10.1016/j.cma.2024.116813_b42) 2019; 62
Vadyala (10.1016/j.cma.2024.116813_b49) 2022; 13
Lagaris (10.1016/j.cma.2024.116813_b28) 1998; 9
Wang (10.1016/j.cma.2024.116813_b40) 2021
Jagtap (10.1016/j.cma.2024.116813_b21) 2020; 28
Mattey (10.1016/j.cma.2024.116813_b25) 2022; 390
Liang (10.1016/j.cma.2024.116813_b20) 2021
Wight (10.1016/j.cma.2024.116813_b16) 2020
Schaarschmidt (10.1016/j.cma.2024.116813_b60) 2021
Lorenz (10.1016/j.cma.2024.116813_b56) 1963; 20
References_xml – year: 2021
  ident: b20
  article-title: Reproducing activation function for deep learning
– volume: 28
  start-page: 11618
  year: 2020
  end-page: 11633
  ident: b7
  article-title: Physics-informed neural networks for inverse problems in nano-optics and metamaterials
  publication-title: Opt. Express
– volume: 404
  year: 2020
  ident: b23
  article-title: Adaptive activation functions accelerate convergence in deep and physics-informed neural networks
  publication-title: J. Comput. Phys.
– year: 2021
  ident: b17
  article-title: Efficient training of physics-informed neural networks via importance sampling
  publication-title: Comput.-Aided Civ. Infrastruct. Eng.
– volume: 449
  year: 2022
  ident: b13
  article-title: When and why PINNs fail to train: A neural tangent kernel perspective
  publication-title: J. Comput. Phys.
– year: 2021
  ident: b41
  article-title: Physics-informed neural operator for learning partial differential equations
– year: 2020
  ident: b46
  article-title: Tilted empirical risk minimization
– year: 2009
  ident: b33
  article-title: A First Course in the Numerical Analysis of Differential Equations
– year: 2014
  ident: b32
  article-title: Adam: A method for stochastic optimization
– year: 2022
  ident: b44
  article-title: Causal machine learning: A survey and open problems
– year: 2021
  ident: b60
  article-title: Automap: Towards ergonomic automated parallelism for ML models
– volume: 118
  year: 2021
  ident: b62
  article-title: Machine learning–accelerated computational fluid dynamics
  publication-title: Proc. Natl. Acad. Sci.
– year: 2020
  ident: b14
  article-title: Self-adaptive physics-informed neural networks using a soft attention mechanism
– volume: 384
  year: 2021
  ident: b9
  article-title: On the eigenvector bias of Fourier feature networks: From regression to solving multi-scale PDEs with physics-informed neural networks
  publication-title: Comput. Methods Appl. Mech. Engrg.
– year: 2021
  ident: b51
  article-title: Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks
– volume: 28
  start-page: 2002
  year: 2020
  end-page: 2041
  ident: b21
  article-title: Extended physics-informed neural networks (XPINNs): A generalized space-time domain decomposition based deep learning framework for nonlinear partial differential equations
  publication-title: Commun. Comput. Phys.
– year: 2019
  ident: b36
  article-title: DeepXDE: A deep learning library for solving differential equations
– volume: 19
  start-page: 932
  year: 2018
  end-page: 955
  ident: b59
  article-title: Deep hidden physics models: Deep learning of nonlinear partial differential equations
  publication-title: J. Mach. Learn. Res.
– year: 2021
  ident: b38
  article-title: Learning the solution operator of parametric partial differential equations with physics-informed DeepOnets
– start-page: 872
  year: 2019
  end-page: 881
  ident: b47
  article-title: What is the effect of importance weighting in deep learning?
  publication-title: International Conference on Machine Learning
– year: 2008
  ident: b31
  article-title: Evaluating Derivatives: Principles and Techniques of Algorithmic Differentiation
– volume: 39
  start-page: 1
  year: 2020
  end-page: 20
  ident: b6
  article-title: Physics-informed neural network for ultrasound nondestructive quantification of surface breaking cracks
  publication-title: J. Nondestruct. Eval.
– year: 1998
  ident: b27
  publication-title: Partial Differential Equations
– volume: 468
  start-page: 165
  year: 2022
  end-page: 180
  ident: b19
  article-title: Deep Kronecker neural networks: A general framework for neural networks with adaptive activation functions
  publication-title: Neurocomputing
– volume: 358
  year: 2020
  ident: b3
  article-title: Machine learning in cardiovascular flows modeling: Predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
  publication-title: Comput. Methods Appl. Mech. Engrg.
– volume: 38
  start-page: 1499
  year: 1992
  end-page: 1511
  ident: b54
  article-title: A hybrid neural network-first principles approach to process modeling
  publication-title: AIChE J.
– volume: 13
  year: 2022
  ident: b49
  article-title: Physics-informed neural network method for solving one-dimensional advection equation using PyTorch
  publication-title: Array
– volume: 428
  year: 2021
  ident: b5
  article-title: Deep learning of free boundary and Stefan problems
  publication-title: J. Comput. Phys.
– reference: Xavier Glorot, Yoshua Bengio, Understanding the difficulty of training deep feedforward neural networks, in: Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 2010, pp. 249–256.
– volume: 4
  start-page: 1177
  year: 1977
  end-page: 1206
  ident: b58
  article-title: Nonlinear analysis of hydrodynamic instability in laminar flames—I. derivation of basic equations
  publication-title: Acta Astronaut.
– start-page: 675
  year: 2021
  end-page: 683
  ident: b18
  article-title: Quadratic residual networks: A new class of neural networks for solving forward and inverse problems in physics involving PDEs
  publication-title: Proceedings of the 2021 SIAM International Conference on Data Mining
– volume: 9
  start-page: 987
  year: 1998
  end-page: 1000
  ident: b28
  article-title: Artificial neural networks for solving ordinary and partial differential equations
  publication-title: IEEE Trans. Neural Netw.
– volume: vol. 33
  start-page: 12697
  year: 2020
  end-page: 12709
  ident: b45
  article-title: Learning causal effects via weighted empirical risk minimization
  publication-title: Advances in Neural Information Processing Systems
– volume: 104
  year: 2021
  ident: b2
  article-title: Uncovering turbulent plasma dynamics via deep learning from partial observations
  publication-title: Phys. Rev. E
– year: 2021
  ident: b48
  article-title: Evolutional deep neural network
– volume: 43
  start-page: A3055
  year: 2021
  end-page: A3081
  ident: b12
  article-title: Understanding and mitigating gradient flow pathologies in physics-informed neural networks
  publication-title: SIAM J. Sci. Comput.
– year: 2021
  ident: b61
  article-title: Parallel physics-informed neural networks via domain decomposition
– volume: 20
  start-page: 130
  year: 1963
  end-page: 141
  ident: b56
  article-title: Deterministic nonperiodic flow
  publication-title: J. Atmos. Sci.
– volume: 62
  start-page: 54
  year: 2019
  end-page: 60
  ident: b42
  article-title: The seven tools of causal inference, with reflections on machine learning
  publication-title: Commun. ACM
– year: 2021
  ident: b52
  article-title: Physics-informed neural networks with hard constraints for inverse design
– volume: 378
  start-page: 686
  year: 2019
  end-page: 707
  ident: b11
  article-title: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
  publication-title: J. Comput. Phys.
– year: 2018
  ident: b35
  article-title: JAX: composable transformations of python+numpy programs
– start-page: 447
  year: 2021
  end-page: 461
  ident: b37
  article-title: NVIDIA SimNet™: An AI-accelerated multi-physics simulation framework
  publication-title: International Conference on Computational Science
– volume: 390
  year: 2022
  ident: b25
  article-title: A novel sequential method to train physics informed neural networks for Allen Cahn and Cahn Hilliard equations
  publication-title: Comput. Methods Appl. Mech. Engrg.
– volume: 55
  start-page: 356
  year: 1976
  end-page: 369
  ident: b57
  article-title: Persistent propagation of concentration waves in dissipative media far from thermal equilibrium
  publication-title: Prog. Theor. Phys.
– volume: 16
  year: 2020
  ident: b4
  article-title: Systems biology informed deep learning for inferring parameters and hidden dynamics
  publication-title: PLoS Comput. Biol.
– volume: 9
  start-page: 90
  year: 2007
  end-page: 95
  ident: b63
  article-title: Matplotlib: A 2D graphics environment
  publication-title: IEEE Ann. Hist. Comput.
– volume: 367
  start-page: 1026
  year: 2020
  end-page: 1030
  ident: b1
  article-title: Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
  publication-title: Science
– year: 2021
  ident: b39
  article-title: Long-time integration of parametric evolution equations with physics-informed DeepONets
– volume: 4
  year: 2019
  ident: b29
  article-title: Deep learning of turbulent scalar mixing
  publication-title: Phys. Rev. Fluids
– start-page: 8571
  year: 2018
  end-page: 8580
  ident: b34
  article-title: Neural tangent kernel: Convergence and generalization in neural networks
  publication-title: Advances in Neural Information Processing Systems
– year: 2007
  ident: b26
  article-title: Partial Differential Equations: An Introduction
– year: 2019
  ident: b30
  article-title: Variational physics-informed neural networks for solving partial differential equations
– year: 2019
  ident: b53
  article-title: Taylor-mode automatic differentiation for higher-order derivatives in JAX
– year: 2020
  ident: b16
  article-title: Solving Allen-Cahn and Cahn-Hilliard equations using the adaptive physics informed neural networks
– year: 2021
  ident: b40
  article-title: Improved architectures and training algorithms for deep operator networks
– volume: 8
  start-page: 42
  year: 2020
  ident: b8
  article-title: Physics-informed neural networks for cardiac activation mapping
  publication-title: Front. Phys.
– year: 2021
  ident: b24
  article-title: Characterizing possible failure modes in physics-informed neural networks
– volume: 435
  year: 2021
  ident: b50
  article-title: A method for representing periodic functions and enforcing exactly periodic boundary conditions with deep neural networks
  publication-title: J. Comput. Phys.
– year: 2021
  ident: b15
  article-title: Inverse Dirichlet weighting enables reliable training of physics informed neural networks
  publication-title: Mach. Learn.: Sci. Technol.
– volume: 585
  start-page: 357
  year: 2020
  end-page: 362
  ident: b64
  article-title: Array programming with NumPy
  publication-title: Nature
– start-page: 1
  year: 2021
  end-page: 19
  ident: b10
  article-title: Physics-informed machine learning
  publication-title: Nat. Rev. Phys.
– year: 2021
  ident: b22
  article-title: Finite basis physics-informed neural networks (FBPINNs): a scalable domain decomposition approach for solving differential equations
– volume: 109
  start-page: 612
  year: 2021
  end-page: 634
  ident: b43
  article-title: Toward causal representation learning
  publication-title: Proc. IEEE
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b39
– volume: 358
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b3
  article-title: Machine learning in cardiovascular flows modeling: Predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
  publication-title: Comput. Methods Appl. Mech. Engrg.
  doi: 10.1016/j.cma.2019.112623
– ident: 10.1016/j.cma.2024.116813_b55
– start-page: 447
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b37
  article-title: NVIDIA SimNet™: An AI-accelerated multi-physics simulation framework
– volume: 468
  start-page: 165
  year: 2022
  ident: 10.1016/j.cma.2024.116813_b19
  article-title: Deep Kronecker neural networks: A general framework for neural networks with adaptive activation functions
  publication-title: Neurocomputing
  doi: 10.1016/j.neucom.2021.10.036
– volume: 43
  start-page: A3055
  issue: 5
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b12
  article-title: Understanding and mitigating gradient flow pathologies in physics-informed neural networks
  publication-title: SIAM J. Sci. Comput.
  doi: 10.1137/20M1318043
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b61
– year: 1998
  ident: 10.1016/j.cma.2024.116813_b27
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b24
– volume: 13
  year: 2022
  ident: 10.1016/j.cma.2024.116813_b49
  article-title: Physics-informed neural network method for solving one-dimensional advection equation using PyTorch
  publication-title: Array
  doi: 10.1016/j.array.2021.100110
– volume: 585
  start-page: 357
  issue: 7825
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b64
  article-title: Array programming with NumPy
  publication-title: Nature
  doi: 10.1038/s41586-020-2649-2
– start-page: 8571
  year: 2018
  ident: 10.1016/j.cma.2024.116813_b34
  article-title: Neural tangent kernel: Convergence and generalization in neural networks
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b15
  article-title: Inverse Dirichlet weighting enables reliable training of physics informed neural networks
  publication-title: Mach. Learn.: Sci. Technol.
– volume: 109
  start-page: 612
  issue: 5
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b43
  article-title: Toward causal representation learning
  publication-title: Proc. IEEE
  doi: 10.1109/JPROC.2021.3058954
– year: 2007
  ident: 10.1016/j.cma.2024.116813_b26
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b41
– volume: 367
  start-page: 1026
  issue: 6481
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b1
  article-title: Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
  publication-title: Science
  doi: 10.1126/science.aaw4741
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b52
– volume: 8
  start-page: 42
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b8
  article-title: Physics-informed neural networks for cardiac activation mapping
  publication-title: Front. Phys.
  doi: 10.3389/fphy.2020.00042
– volume: 449
  year: 2022
  ident: 10.1016/j.cma.2024.116813_b13
  article-title: When and why PINNs fail to train: A neural tangent kernel perspective
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2021.110768
– year: 2019
  ident: 10.1016/j.cma.2024.116813_b53
– year: 2020
  ident: 10.1016/j.cma.2024.116813_b16
– volume: 9
  start-page: 90
  issue: 03
  year: 2007
  ident: 10.1016/j.cma.2024.116813_b63
  article-title: Matplotlib: A 2D graphics environment
  publication-title: IEEE Ann. Hist. Comput.
– year: 2020
  ident: 10.1016/j.cma.2024.116813_b14
– volume: 435
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b50
  article-title: A method for representing periodic functions and enforcing exactly periodic boundary conditions with deep neural networks
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2021.110242
– volume: 4
  start-page: 1177
  issue: 11
  year: 1977
  ident: 10.1016/j.cma.2024.116813_b58
  article-title: Nonlinear analysis of hydrodynamic instability in laminar flames—I. derivation of basic equations
  publication-title: Acta Astronaut.
  doi: 10.1016/0094-5765(77)90096-0
– start-page: 1
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b10
  article-title: Physics-informed machine learning
  publication-title: Nat. Rev. Phys.
– volume: 62
  start-page: 54
  issue: 3
  year: 2019
  ident: 10.1016/j.cma.2024.116813_b42
  article-title: The seven tools of causal inference, with reflections on machine learning
  publication-title: Commun. ACM
  doi: 10.1145/3241036
– volume: 20
  start-page: 130
  issue: 2
  year: 1963
  ident: 10.1016/j.cma.2024.116813_b56
  article-title: Deterministic nonperiodic flow
  publication-title: J. Atmos. Sci.
  doi: 10.1175/1520-0469(1963)020<0130:DNF>2.0.CO;2
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b20
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b22
– volume: 28
  start-page: 11618
  issue: 8
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b7
  article-title: Physics-informed neural networks for inverse problems in nano-optics and metamaterials
  publication-title: Opt. Express
  doi: 10.1364/OE.384875
– year: 2008
  ident: 10.1016/j.cma.2024.116813_b31
– year: 2022
  ident: 10.1016/j.cma.2024.116813_b44
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b40
– volume: 28
  start-page: 2002
  issue: 5
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b21
  article-title: Extended physics-informed neural networks (XPINNs): A generalized space-time domain decomposition based deep learning framework for nonlinear partial differential equations
  publication-title: Commun. Comput. Phys.
  doi: 10.4208/cicp.OA-2020-0164
– volume: 404
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b23
  article-title: Adaptive activation functions accelerate convergence in deep and physics-informed neural networks
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2019.109136
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b17
  article-title: Efficient training of physics-informed neural networks via importance sampling
  publication-title: Comput.-Aided Civ. Infrastruct. Eng.
  doi: 10.1111/mice.12685
– volume: 38
  start-page: 1499
  issue: 10
  year: 1992
  ident: 10.1016/j.cma.2024.116813_b54
  article-title: A hybrid neural network-first principles approach to process modeling
  publication-title: AIChE J.
  doi: 10.1002/aic.690381003
– volume: 55
  start-page: 356
  issue: 2
  year: 1976
  ident: 10.1016/j.cma.2024.116813_b57
  article-title: Persistent propagation of concentration waves in dissipative media far from thermal equilibrium
  publication-title: Prog. Theor. Phys.
  doi: 10.1143/PTP.55.356
– start-page: 872
  year: 2019
  ident: 10.1016/j.cma.2024.116813_b47
  article-title: What is the effect of importance weighting in deep learning?
– volume: 19
  start-page: 932
  issue: 1
  year: 2018
  ident: 10.1016/j.cma.2024.116813_b59
  article-title: Deep hidden physics models: Deep learning of nonlinear partial differential equations
  publication-title: J. Mach. Learn. Res.
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b38
– volume: 104
  issue: 2
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b2
  article-title: Uncovering turbulent plasma dynamics via deep learning from partial observations
  publication-title: Phys. Rev. E
  doi: 10.1103/PhysRevE.104.025205
– volume: 9
  start-page: 987
  issue: 5
  year: 1998
  ident: 10.1016/j.cma.2024.116813_b28
  article-title: Artificial neural networks for solving ordinary and partial differential equations
  publication-title: IEEE Trans. Neural Netw.
  doi: 10.1109/72.712178
– volume: 4
  issue: 12
  year: 2019
  ident: 10.1016/j.cma.2024.116813_b29
  article-title: Deep learning of turbulent scalar mixing
  publication-title: Phys. Rev. Fluids
  doi: 10.1103/PhysRevFluids.4.124501
– year: 2019
  ident: 10.1016/j.cma.2024.116813_b30
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b60
– year: 2019
  ident: 10.1016/j.cma.2024.116813_b36
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b51
– year: 2018
  ident: 10.1016/j.cma.2024.116813_b35
– volume: vol. 33
  start-page: 12697
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b45
  article-title: Learning causal effects via weighted empirical risk minimization
– volume: 428
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b5
  article-title: Deep learning of free boundary and Stefan problems
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2020.109914
– year: 2020
  ident: 10.1016/j.cma.2024.116813_b46
– volume: 378
  start-page: 686
  year: 2019
  ident: 10.1016/j.cma.2024.116813_b11
  article-title: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
  publication-title: J. Comput. Phys.
  doi: 10.1016/j.jcp.2018.10.045
– year: 2014
  ident: 10.1016/j.cma.2024.116813_b32
– volume: 384
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b9
  article-title: On the eigenvector bias of Fourier feature networks: From regression to solving multi-scale PDEs with physics-informed neural networks
  publication-title: Comput. Methods Appl. Mech. Engrg.
  doi: 10.1016/j.cma.2021.113938
– year: 2021
  ident: 10.1016/j.cma.2024.116813_b48
– volume: 390
  year: 2022
  ident: 10.1016/j.cma.2024.116813_b25
  article-title: A novel sequential method to train physics informed neural networks for Allen Cahn and Cahn Hilliard equations
  publication-title: Comput. Methods Appl. Mech. Engrg.
  doi: 10.1016/j.cma.2021.114474
– volume: 39
  start-page: 1
  issue: 3
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b6
  article-title: Physics-informed neural network for ultrasound nondestructive quantification of surface breaking cracks
  publication-title: J. Nondestruct. Eval.
  doi: 10.1007/s10921-020-00705-1
– start-page: 675
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b18
  article-title: Quadratic residual networks: A new class of neural networks for solving forward and inverse problems in physics involving PDEs
– year: 2009
  ident: 10.1016/j.cma.2024.116813_b33
– volume: 16
  issue: 11
  year: 2020
  ident: 10.1016/j.cma.2024.116813_b4
  article-title: Systems biology informed deep learning for inferring parameters and hidden dynamics
  publication-title: PLoS Comput. Biol.
  doi: 10.1371/journal.pcbi.1007575
– volume: 118
  issn: 0027-8424
  issue: 21
  year: 2021
  ident: 10.1016/j.cma.2024.116813_b62
  article-title: Machine learning–accelerated computational fluid dynamics
  publication-title: Proc. Natl. Acad. Sci.
  doi: 10.1073/pnas.2101784118
SSID ssj0000812
Score 2.718097
Snippet While the popularity of physics-informed neural networks (PINNs) is steadily rising, to this date PINNs have not been successful in simulating dynamical...
SourceID osti
crossref
elsevier
SourceType Open Access Repository
Enrichment Source
Index Database
Publisher
StartPage 116813
SubjectTerms Chaotic systems
Computational physics
Deep learning
Partial differential equations
Title Respecting causality for training physics-informed neural networks
URI https://dx.doi.org/10.1016/j.cma.2024.116813
https://www.osti.gov/biblio/2290430
Volume 421
WOSCitedRecordID wos001179213600001&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: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  customDbUrl:
  eissn: 1879-2138
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0000812
  issn: 0045-7825
  databaseCode: AIEXJ
  dateStart: 19950101
  isFulltext: true
  titleUrlDefault: https://www.sciencedirect.com
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1NbxMxELUg5QAHPgqIUkA-cEIy2g87to-lKgIOFYIi5bayvbZIS7dVNkXl3zNej7MhgggOXFbRJl6tZl5mn2dn3hDy0payBSLhmW6FY9xMBVNeBibbKO82rX0p7DBsQh4fq9lMf8RS3n4YJyC7Tl1f68v_6mo4B86OrbP_4O7VReEEfAanwxHcDse_cvwnn7onYzOtueoTzx6KCXEaBGYzepZEU-P7fz-Ib3SpJLxfJ6x56gOOmh6qZw0y13Mf24azzLMflQ3HLD0mo-dhBOFn052ZBeZdv_4w52N0BrCexaGIidou5r_kJCo-FmWlRBk-1dcDLxcMyEh6ge1TrFVSs6pM4i45GPPUL42oO_xtkE_5htPXbhCOqjiE_alKHa0b2tlRzJ7XxU2yU0mhIejtHLw_mn0YH9WqTHLyeHf5tfdQALhx-T8Rl8kFxOI1TnJyn9zFzQQ9SCB4QG74bpfcw40FxbDd75I7a6qTD8mbESF0hRAKUKAZIXQTITQhhGaEPCJf3h6dHL5jOEqDOV7XSyZaDlTOtM4YXbRSW7C9h91nxUOhgi94W3PrAq-VsBooHQ_aSqt5ADYndLD1YzLpLjr_hFAwFvyPXTGtvOeKFybISoQKgjm3OpT1HimynRqHOvPx9r81uaDwtAHTNtG0TTLtHnm1WnKZRFa2_Zhn4zfIEhP7awAf25btR0fFJVEd2cUyMliDCHm69dt9cntE-TMyWS6u_HNyy31fzvvFCwTVTw33jwc
linkProvider Elsevier
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=Respecting+causality+for+training+physics-informed+neural+networks&rft.jtitle=Computer+methods+in+applied+mechanics+and+engineering&rft.au=Wang%2C+Sifan&rft.au=Sankaran%2C+Shyam&rft.au=Perdikaris%2C+Paris&rft.date=2024-03-01&rft.pub=Elsevier&rft.issn=0045-7825&rft.eissn=1879-2138&rft.volume=421&rft.issue=C&rft_id=info:doi/10.1016%2Fj.cma.2024.116813&rft.externalDocID=2290430
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0045-7825&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0045-7825&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0045-7825&client=summon