Skip to main content
Erschienen in:

10.04.2024 | Original Article

NeuFENet: neural finite element solutions with theoretical bounds for parametric PDEs

verfasst von: Biswajit Khara, Aditya Balu, Ameya Joshi, Soumik Sarkar, Chinmay Hegde, Adarsh Krishnamurthy, Baskar Ganapathysubramanian

Erschienen in: Engineering with Computers | Ausgabe 5/2024

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config
loading …

Abstract

We consider a mesh-based approach for training a neural network to produce field predictions of solutions to parametric partial differential equations (PDEs). This approach contrasts current approaches for “neural PDE solvers” that employ collocation-based methods to make pointwise predictions of solutions to PDEs. This approach has the advantage of naturally enforcing different boundary conditions as well as ease of invoking well-developed PDE theory—including analysis of numerical stability and convergence—to obtain capacity bounds for our proposed neural networks in discretized domains. We explore our mesh-based strategy, called NeuFENet, using a weighted Galerkin loss function based on the Finite Element Method (FEM) on a parametric elliptic PDE. The weighted Galerkin loss (FEM loss) is similar to an energy functional that produces improved solutions, satisfies a priori mesh convergence, and can model Dirichlet and Neumann boundary conditions. We prove theoretically, and illustrate with experiments, convergence results analogous to mesh convergence analysis deployed in finite element solutions to PDEs. These results suggest that a mesh-based neural network approach serves as a promising approach for solving parametric PDEs with theoretical bounds.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 390 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Maschinenbau + Werkstoffe




 

Jetzt Wissensvorsprung sichern!

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

  • über 102.000 Bücher
  • über 537 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Maschinenbau + Werkstoffe
  • Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 340 Zeitschriften

aus folgenden Fachgebieten:

  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Versicherung + Risiko




Jetzt Wissensvorsprung sichern!

Anhänge
Nur mit Berechtigung zugänglich
Fußnoten
1
In contrast, state-of-art neural methods allow us to use basis functions beyond polynomials or Fourier bases and approximate much more complicated mappings. Although such methods can be analyzed theoretically, the estimates are often impractical [2527] This is a very active area of research, and we expect tighter estimates in the future.
 
2
While a probability-based definition of \(\omega\) is not needed for defining a parameteric PDE, we choose this definition for two reasons. First, such a formulation allows easy extension to the stochastic PDE case. Second, such a formulation will allow using expectation-based arguments in the analysis of convergence.
 
Literatur
1.
Zurück zum Zitat Raissi M, Perdikaris P, Karniadakis GE (2019) Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J Comput Phys 378:686–707MathSciNetCrossRef Raissi M, Perdikaris P, Karniadakis GE (2019) Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J Comput Phys 378:686–707MathSciNetCrossRef
2.
Zurück zum Zitat Rudy S, Alla A, Brunton SL, Nathan Kutz J (2019) Data-driven identification of parametric partial differential equations. SIAM J Appl Dyna Syst 18(2):643–660MathSciNetCrossRef Rudy S, Alla A, Brunton SL, Nathan Kutz J (2019) Data-driven identification of parametric partial differential equations. SIAM J Appl Dyna Syst 18(2):643–660MathSciNetCrossRef
3.
Zurück zum Zitat Tompson J, Schlachter K, Sprechmann P, Perlin K (2017)Accelerating Eulerian fluid simulation with convolutional networks. In: International Conference on machine learning, pp 3424–3433. PMLR, Tompson J, Schlachter K, Sprechmann P, Perlin K (2017)Accelerating Eulerian fluid simulation with convolutional networks. In: International Conference on machine learning, pp 3424–3433. PMLR,
4.
Zurück zum Zitat Raissi M, Karniadakis GE (2018) Hidden physics models: machine learning of nonlinear partial differential equations. J Comput Phys 357:125–141MathSciNetCrossRef Raissi M, Karniadakis GE (2018) Hidden physics models: machine learning of nonlinear partial differential equations. J Comput Phys 357:125–141MathSciNetCrossRef
5.
Zurück zum Zitat Lu L, Jin P, Karniadakis GE (2019) Deeponet: learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators. arXiv preprint arXiv:1910.03193, Lu L, Jin P, Karniadakis GE (2019) Deeponet: learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators. arXiv preprint arXiv:​1910.​03193,
6.
Zurück zum Zitat Kharazmi E, Zhang Z, Karniadakis GE (2019) Variational physics-informed neural networks for solving partial differential equations. arXiv preprint arXiv:1912.00873, Kharazmi E, Zhang Z, Karniadakis GE (2019) Variational physics-informed neural networks for solving partial differential equations. arXiv preprint arXiv:​1912.​00873,
7.
Zurück zum Zitat Sirignano J, Spiliopoulos K (2018) Dgm: a deep learning algorithm for solving partial differential equations. J Comput Phys 375:1339–1364MathSciNetCrossRef Sirignano J, Spiliopoulos K (2018) Dgm: a deep learning algorithm for solving partial differential equations. J Comput Phys 375:1339–1364MathSciNetCrossRef
8.
Zurück zum Zitat Yang L, Dongkun Z, Karniadakis GE (2018) Physics-informed generative adversarial networks for stochastic differential equations. arXiv preprint arXiv:1811.02033, Yang L, Dongkun Z, Karniadakis GE (2018) Physics-informed generative adversarial networks for stochastic differential equations. arXiv preprint arXiv:​1811.​02033,
9.
Zurück zum Zitat Guofei Pang LL, Karniadakis GE (2019) fpinns: fractional physics-informed neural networks. SIAM J Sci Comput 41(4):A2603–A2626MathSciNetCrossRef Guofei Pang LL, Karniadakis GE (2019) fpinns: fractional physics-informed neural networks. SIAM J Sci Comput 41(4):A2603–A2626MathSciNetCrossRef
10.
Zurück zum Zitat Karumuri S, Tripathy R, Bilionis I, Panchal J (2020) Simulator-free solution of high-dimensional stochastic elliptic partial differential equations using deep neural networks. J Comput Phys 404:109120MathSciNetCrossRef Karumuri S, Tripathy R, Bilionis I, Panchal J (2020) Simulator-free solution of high-dimensional stochastic elliptic partial differential equations using deep neural networks. J Comput Phys 404:109120MathSciNetCrossRef
11.
Zurück zum Zitat Han J, Arnulf J, Weinan E (2018) Solving high-dimensional partial differential equations using deep learning. Proc Natl Acad Sci 115(34):8505–8510MathSciNetCrossRef Han J, Arnulf J, Weinan E (2018) Solving high-dimensional partial differential equations using deep learning. Proc Natl Acad Sci 115(34):8505–8510MathSciNetCrossRef
12.
Zurück zum Zitat Michoski C, Milosavljevic M, Oliver T, Hatch D (2019) Solving irregular and data-enriched differential equations using deep neural networks. arXiv preprint arXiv:1905.04351 Michoski C, Milosavljevic M, Oliver T, Hatch D (2019) Solving irregular and data-enriched differential equations using deep neural networks. arXiv preprint arXiv:​1905.​04351
13.
Zurück zum Zitat Samaniego E, Anitescu C, Goswami S, Nguyen-Thanh VM, Hongwei G, Khader H, Zhuang X, Rabczuk T (2020) An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications. Comput Methods Appl Mech Eng 362:112790MathSciNetCrossRef Samaniego E, Anitescu C, Goswami S, Nguyen-Thanh VM, Hongwei G, Khader H, Zhuang X, Rabczuk T (2020) An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications. Comput Methods Appl Mech Eng 362:112790MathSciNetCrossRef
14.
Zurück zum Zitat Ramabathiran Amuthan A, Prabhu R (2021) Spinn: sparse, physics-based, and partially interpretable neural networks for pdes. J Comput Phys 445:110600MathSciNetCrossRef Ramabathiran Amuthan A, Prabhu R (2021) Spinn: sparse, physics-based, and partially interpretable neural networks for pdes. J Comput Phys 445:110600MathSciNetCrossRef
15.
Zurück zum Zitat Botelho S, Joshi A, Khara B, Sarkar S, Hegde C, Adavani S, Ganapathysubramanian B (2020) Deep generative models that solve pdes: distributed computing for training large data-free models. arXiv preprint arXiv:2007.12792, Botelho S, Joshi A, Khara B, Sarkar S, Hegde C, Adavani S, Ganapathysubramanian B (2020) Deep generative models that solve pdes: distributed computing for training large data-free models. arXiv preprint arXiv:​2007.​12792,
16.
Zurück zum Zitat Mitusch Sebastian K, Funke Simon W, Miroslav K (2021) Hybrid fem-nn models: combining artificial neural networks with the finite element method. J Comput Phys 446:110651MathSciNetCrossRef Mitusch Sebastian K, Funke Simon W, Miroslav K (2021) Hybrid fem-nn models: combining artificial neural networks with the finite element method. J Comput Phys 446:110651MathSciNetCrossRef
17.
Zurück zum Zitat Jokar M, Semperlotti F (2021) Finite element network analysis: a machine learning based computational framework for the simulation of physical systems. Comput Struct 247:106484CrossRef Jokar M, Semperlotti F (2021) Finite element network analysis: a machine learning based computational framework for the simulation of physical systems. Comput Struct 247:106484CrossRef
18.
Zurück zum Zitat Sitzmann V, Martel JNP, Bergman AW, Lindell DB, Wetzstein G (2020) Implicit neural representations with periodic activation functions. Adv Neural Inform Process Syst 33:7462–7473 Sitzmann V, Martel JNP, Bergman AW, Lindell DB, Wetzstein G (2020) Implicit neural representations with periodic activation functions. Adv Neural Inform Process Syst 33:7462–7473
19.
Zurück zum Zitat Mishra S, Konstantin Rusch T (2021) Enhancing accuracy of deep learning algorithms by training with low-discrepancy sequences. SIAM J Numer Anal 59(3):1811–1834MathSciNetCrossRef Mishra S, Konstantin Rusch T (2021) Enhancing accuracy of deep learning algorithms by training with low-discrepancy sequences. SIAM J Numer Anal 59(3):1811–1834MathSciNetCrossRef
20.
Zurück zum Zitat Zhu Y, Zabaras N, Koutsourelakis P-S, Perdikaris P (2019) Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data. J Comput Phys 394:56–81MathSciNetCrossRef Zhu Y, Zabaras N, Koutsourelakis P-S, Perdikaris P (2019) Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data. J Comput Phys 394:56–81MathSciNetCrossRef
21.
Zurück zum Zitat Wen G, Li Z, Azizzadenesheli K, Anandkumar A, Benson Sally M (2021) U-fno–an enhanced fourier neural operator based-deep learning model for multiphase flow. arXiv preprint arXiv:2109.03697, Wen G, Li Z, Azizzadenesheli K, Anandkumar A, Benson Sally M (2021) U-fno–an enhanced fourier neural operator based-deep learning model for multiphase flow. arXiv preprint arXiv:​2109.​03697,
22.
Zurück zum Zitat Ranade R, Hill C, Pathak J (2021) Discretizationnet: a machine-learning based solver for Navier-Stokes equations using finite volume discretization. Comput Methods Appl Mech Eng 378:113722MathSciNetCrossRef Ranade R, Hill C, Pathak J (2021) Discretizationnet: a machine-learning based solver for Navier-Stokes equations using finite volume discretization. Comput Methods Appl Mech Eng 378:113722MathSciNetCrossRef
23.
Zurück zum Zitat Brenner S, Scott R (2007) The mathematical theory of finite element methods, vol 15. Springer Science & Business Media Brenner S, Scott R (2007) The mathematical theory of finite element methods, vol 15. Springer Science & Business Media
24.
Zurück zum Zitat Larson MG, Bengzon F (2013) The finite element method: theory, implementation, and applications, vol 10. Springer Science & Business MediaCrossRef Larson MG, Bengzon F (2013) The finite element method: theory, implementation, and applications, vol 10. Springer Science & Business MediaCrossRef
25.
Zurück zum Zitat Shin Y , Zhang Z, Karniadakis GE (2020) Error estimates of residual minimization using neural networks for linear pdes. arXiv preprint arXiv:2010.08019, Shin Y , Zhang Z, Karniadakis GE (2020) Error estimates of residual minimization using neural networks for linear pdes. arXiv preprint arXiv:​2010.​08019,
26.
Zurück zum Zitat Mishra S, Molinaro R (2020) Estimates on the generalization error of physics informed neural networks (pinns) for approximating pdes. arXiv preprint arXiv:2006.16144 Mishra S, Molinaro R (2020) Estimates on the generalization error of physics informed neural networks (pinns) for approximating pdes. arXiv preprint arXiv:​2006.​16144
27.
Zurück zum Zitat Jiao Y, Lai Y, Luo Y, Wang Y, Yang Y (2021) Error analysis of deep ritz methods for elliptic equations. arXiv preprint arXiv:2107.14478 Jiao Y, Lai Y, Luo Y, Wang Y, Yang Y (2021) Error analysis of deep ritz methods for elliptic equations. arXiv preprint arXiv:​2107.​14478
28.
29.
Zurück zum Zitat Takeuchi J, Kosugi Y (1994) Neural network representation of finite element method. Neural Netw 7(2):389–395CrossRef Takeuchi J, Kosugi Y (1994) Neural network representation of finite element method. Neural Netw 7(2):389–395CrossRef
30.
Zurück zum Zitat Xu G, Littlefair G, Penson R, Callan R (1999) Application of fe-based neural networks to dynamic problems. In: ICONIP’99. ANZIIS’99 & ANNES’99 & ACNN’99. 6th International Conference on neural information processing. Proceedings (Cat. No. 99EX378), volume 3, pp 1039–1044. IEEE Xu G, Littlefair G, Penson R, Callan R (1999) Application of fe-based neural networks to dynamic problems. In: ICONIP’99. ANZIIS’99 & ANNES’99 & ACNN’99. 6th International Conference on neural information processing. Proceedings (Cat. No. 99EX378), volume 3, pp 1039–1044. IEEE
31.
Zurück zum Zitat Ramuhalli P, Udpa L, Udpa SS (2005) Finite-element neural networks for solving differential equations. IEEE Trans Neural Netw 16(6):1381–1392CrossRef Ramuhalli P, Udpa L, Udpa SS (2005) Finite-element neural networks for solving differential equations. IEEE Trans Neural Netw 16(6):1381–1392CrossRef
32.
Zurück zum Zitat Chao X, Wang C, Ji F, Yuan X (2012) Finite-element neural network-based solving 3-d differential equations in mfl. IEEE Trans Magn 48(12):4747–4756CrossRef Chao X, Wang C, Ji F, Yuan X (2012) Finite-element neural network-based solving 3-d differential equations in mfl. IEEE Trans Magn 48(12):4747–4756CrossRef
33.
Zurück zum Zitat Khodayi-Mehr R, Zavlanos M (2020) Varnet: variational neural networks for the solution of partial differential equations. In: Learning for dynamics and control, pp 98–307. PMLR, Khodayi-Mehr R, Zavlanos M (2020) Varnet: variational neural networks for the solution of partial differential equations. In: Learning for dynamics and control, pp 98–307. PMLR,
34.
Zurück zum Zitat Gao H, Zahr Matthew J, Wang J-X (2022) Physics-informed graph neural Galerkin networks: a unified framework for solving pde-governed forward and inverse problems. Comput Methods Appl Mech Eng 390:114502MathSciNetCrossRef Gao H, Zahr Matthew J, Wang J-X (2022) Physics-informed graph neural Galerkin networks: a unified framework for solving pde-governed forward and inverse problems. Comput Methods Appl Mech Eng 390:114502MathSciNetCrossRef
35.
Zurück zum Zitat Yao H, Gao Y, Liu Y (2020) Fea-net: a physics-guided data-driven model for efficient mechanical response prediction. Comput Methods Appl Mech Eng 363:112892MathSciNetCrossRef Yao H, Gao Y, Liu Y (2020) Fea-net: a physics-guided data-driven model for efficient mechanical response prediction. Comput Methods Appl Mech Eng 363:112892MathSciNetCrossRef
36.
Zurück zum Zitat Evans LC (1998) Partial differential equations. Graduate Stud Math 19(4):7 Evans LC (1998) Partial differential equations. Graduate Stud Math 19(4):7
37.
Zurück zum Zitat Reddy JN (2010) An introduction to the finite element method, vol 1221. McGraw-Hill New York Reddy JN (2010) An introduction to the finite element method, vol 1221. McGraw-Hill New York
38.
Zurück zum Zitat Bing Y et al (2017) The deep ritz method: a deep learning-based numerical algorithm for solving variational problems. arXiv preprint arXiv:1710.00211 Bing Y et al (2017) The deep ritz method: a deep learning-based numerical algorithm for solving variational problems. arXiv preprint arXiv:​1710.​00211
39.
Zurück zum Zitat Ming YL et al (2021) Deep Nitsche method: Deep ritz method with essential boundary conditions. Commun Comput Phys 29(5):1365–1384MathSciNetCrossRef Ming YL et al (2021) Deep Nitsche method: Deep ritz method with essential boundary conditions. Commun Comput Phys 29(5):1365–1384MathSciNetCrossRef
41.
Zurück zum Zitat Müller J, Zeinhofer M (2022) Error estimates for the deep ritz method with boundary penalty. In: Mathematical and Scientific Machine Learning, pp 215–230. PMLR Müller J, Zeinhofer M (2022) Error estimates for the deep ritz method with boundary penalty. In: Mathematical and Scientific Machine Learning, pp 215–230. PMLR
42.
Zurück zum Zitat Müller J, Zeinhofer M (2022) Notes on exact boundary values in residual minimisation. In: Mathematical and Scientific Machine Learning, pp 231–240. PMLR Müller J, Zeinhofer M (2022) Notes on exact boundary values in residual minimisation. In: Mathematical and Scientific Machine Learning, pp 231–240. PMLR
43.
Zurück zum Zitat Dondl P, Müller J, Zeinhofer M (2022) Uniform convergence guarantees for the deep ritz method for nonlinear problems. Adv Contin Discrete Models 2022(1):1–19MathSciNet Dondl P, Müller J, Zeinhofer M (2022) Uniform convergence guarantees for the deep ritz method for nonlinear problems. Adv Contin Discrete Models 2022(1):1–19MathSciNet
44.
Zurück zum Zitat Hyuk Lee and In Seok Kang (1990) Neural algorithm for solving differential equations. J Comput Phys 91(1):110–131MathSciNetCrossRef Hyuk Lee and In Seok Kang (1990) Neural algorithm for solving differential equations. J Comput Phys 91(1):110–131MathSciNetCrossRef
45.
Zurück zum Zitat Lagaris IE, Likas A, Fotiadis DI (1998) Artificial neural networks for solving ordinary and partial differential equations. IEEE Trans Neural Netw 9(5):987–1000CrossRef Lagaris IE, Likas A, Fotiadis DI (1998) Artificial neural networks for solving ordinary and partial differential equations. IEEE Trans Neural Netw 9(5):987–1000CrossRef
46.
Zurück zum Zitat Malek A, Beidokhti Shekari R (2006) Numerical solution for high order differential equations using a hybrid neural network optimization method. Appl Math Comput 183(1):260–271MathSciNet Malek A, Beidokhti Shekari R (2006) Numerical solution for high order differential equations using a hybrid neural network optimization method. Appl Math Comput 183(1):260–271MathSciNet
47.
Zurück zum Zitat Sukumar N, Srivastava A (2021) Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks. arXiv preprint arXiv:2104.08426 Sukumar N, Srivastava A (2021) Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks. arXiv preprint arXiv:​2104.​08426
48.
Zurück zum Zitat IsaacE L, AristidisC L, DimitrisG P (2000) Neural-network methods for boundary value problems with irregular boundaries. IEEE Trans Neural Netw 11(5):1041–1049CrossRef IsaacE L, AristidisC L, DimitrisG P (2000) Neural-network methods for boundary value problems with irregular boundaries. IEEE Trans Neural Netw 11(5):1041–1049CrossRef
49.
Zurück zum Zitat Remco Van der M, Cornelis O, Anastasia B (2020) Optimally weighted loss functions for solving pdes with neural networks. arXiv preprint arXiv:2002.06269 Remco Van der M, Cornelis O, Anastasia B (2020) Optimally weighted loss functions for solving pdes with neural networks. arXiv preprint arXiv:​2002.​06269
50.
Zurück zum Zitat Sifan W, Yujun T, and Paris P (2020) Understanding and mitigating gradient pathologies in physics-informed neural networks. arXiv preprint arXiv:2001.04536, Sifan W, Yujun T, and Paris P (2020) Understanding and mitigating gradient pathologies in physics-informed neural networks. arXiv preprint arXiv:​2001.​04536,
51.
Zurück zum Zitat Oliver H, Susheela N, Mohammad Amin N, Akshay S, Kaustubh T, Zhiwei F, Max R, Wonmin B, Sanjay C (2021) Nvidia simnet: an ai-accelerated multi-physics simulation framework. In: International Conference on computational science, pp 447–461. Springer Oliver H, Susheela N, Mohammad Amin N, Akshay S, Kaustubh T, Zhiwei F, Max R, Wonmin B, Sanjay C (2021) Nvidia simnet: an ai-accelerated multi-physics simulation framework. In: International Conference on computational science, pp 447–461. Springer
52.
Zurück zum Zitat Wang S, Perdikaris P (2021) Long-time integration of parametric evolution equations with physics-informed deeponets. arXiv preprint arXiv:2106.05384 Wang S, Perdikaris P (2021) Long-time integration of parametric evolution equations with physics-informed deeponets. arXiv preprint arXiv:​2106.​05384
53.
Zurück zum Zitat Paganini M, de Oliveira L, Nachman B (2018) Calogan: simulating 3d high energy particle showers in multilayer electromagnetic calorimeters with generative adversarial networks. Phys Rev D 97(1):014021CrossRef Paganini M, de Oliveira L, Nachman B (2018) Calogan: simulating 3d high energy particle showers in multilayer electromagnetic calorimeters with generative adversarial networks. Phys Rev D 97(1):014021CrossRef
54.
Zurück zum Zitat Krishnapriyan AS, Gholami A, Zhe S, Kirby RM, Mahoney MW (2021) Characterizing possible failure modes in physics-informed neural networks. arXiv preprint arXiv:2109.01050 Krishnapriyan AS, Gholami A, Zhe S, Kirby RM, Mahoney MW (2021) Characterizing possible failure modes in physics-informed neural networks. arXiv preprint arXiv:​2109.​01050
55.
Zurück zum Zitat Wang S, Teng Y, Perdikaris P (2021) Understanding and mitigating gradient flow pathologies in physics-informed neural networks. SIAM J Sci Comput 43(5):A3055–A3081MathSciNetCrossRef Wang S, Teng Y, Perdikaris P (2021) Understanding and mitigating gradient flow pathologies in physics-informed neural networks. SIAM J Sci Comput 43(5):A3055–A3081MathSciNetCrossRef
56.
Zurück zum Zitat Fox C (1987) An introduction to the calculus of variations. Courier Corporation Fox C (1987) An introduction to the calculus of variations. Courier Corporation
58.
Zurück zum Zitat Olaf R, Philipp F, Brox T (2015) U-net: convolutional networks for biomedical image segmentation. In: International Conference on medical image computing and computer-assisted intervention, pp 234–241. Springer Olaf R, Philipp F, Brox T (2015) U-net: convolutional networks for biomedical image segmentation. In: International Conference on medical image computing and computer-assisted intervention, pp 234–241. Springer
59.
Zurück zum Zitat Özgün Ç, Abdulkadir A, Lienkamp SS, Brox T, Ronneberger O (2016) 3D U-Net: learning dense volumetric segmentation from sparse annotation. In: International Conference on medical image computing and computer-assisted intervention, pp 424–432. Springer Özgün Ç, Abdulkadir A, Lienkamp SS, Brox T, Ronneberger O (2016) 3D U-Net: learning dense volumetric segmentation from sparse annotation. In: International Conference on medical image computing and computer-assisted intervention, pp 424–432. Springer
60.
Zurück zum Zitat Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, Killeen T, Lin Z, Gimelshein N, Antiga L et al (2019) Pytorch: an imperative style, high-performance deep learning library. Adv Neural Inf Process Syst 32:8026–8037 Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, Killeen T, Lin Z, Gimelshein N, Antiga L et al (2019) Pytorch: an imperative style, high-performance deep learning library. Adv Neural Inf Process Syst 32:8026–8037
61.
Zurück zum Zitat Bubeck S, Eldan R, Lee YT, Mikulincer D (2020) Network size and weights size for memorization with two-layers neural networks. arXiv preprint arXiv:2006.02855 Bubeck S, Eldan R, Lee YT, Mikulincer D (2020) Network size and weights size for memorization with two-layers neural networks. arXiv preprint arXiv:​2006.​02855
62.
Zurück zum Zitat LeCun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86(11):2278–2324CrossRef LeCun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86(11):2278–2324CrossRef
63.
Zurück zum Zitat Oden JT, Reddy JN (2012) An introduction to the mathematical theory of finite elements. Courier Corporation Oden JT, Reddy JN (2012) An introduction to the mathematical theory of finite elements. Courier Corporation
64.
Zurück zum Zitat Barron AR (1993) Universal approximation bounds for superpositions of a sigmoidal function. IEEE Trans Inf Theory 39(3):930–945MathSciNetCrossRef Barron AR (1993) Universal approximation bounds for superpositions of a sigmoidal function. IEEE Trans Inf Theory 39(3):930–945MathSciNetCrossRef
65.
Zurück zum Zitat Balu A, Botelho S, Khara B, Rao V, Hegde C, Sarkar S, Adavani S, Krishnamurthy A, Ganapathysubramanian B (2021) Distributed multigrid neural solvers on megavoxel domains. arXiv preprint arXiv:2104.14538 Balu A, Botelho S, Khara B, Rao V, Hegde C, Sarkar S, Adavani S, Krishnamurthy A, Ganapathysubramanian B (2021) Distributed multigrid neural solvers on megavoxel domains. arXiv preprint arXiv:​2104.​14538
66.
Zurück zum Zitat Lu L, Meng X, Mao Z, Karniadakis GE (2021) Deepxde: a deep learning library for solving differential equations. SIAM Rev 63(1):208–228MathSciNetCrossRef Lu L, Meng X, Mao Z, Karniadakis GE (2021) Deepxde: a deep learning library for solving differential equations. SIAM Rev 63(1):208–228MathSciNetCrossRef
68.
Zurück zum Zitat Ghanem RG, Spanos PD (2003) Stochastic finite elements: a spectral approach. Courier Corporation Ghanem RG, Spanos PD (2003) Stochastic finite elements: a spectral approach. Courier Corporation
Metadaten
Titel
NeuFENet: neural finite element solutions with theoretical bounds for parametric PDEs
verfasst von
Biswajit Khara
Aditya Balu
Ameya Joshi
Soumik Sarkar
Chinmay Hegde
Adarsh Krishnamurthy
Baskar Ganapathysubramanian
Publikationsdatum
10.04.2024
Verlag
Springer London
Erschienen in
Engineering with Computers / Ausgabe 5/2024
Print ISSN: 0177-0667
Elektronische ISSN: 1435-5663
DOI
https://doi.org/10.1007/s00366-024-01955-7