fPINNs: Fractional Physics-Informed Neural Networks
Open accessThe DOI matches its Crossref record; 43 of the 43 references it cites were found in a registry or an index.
Abstract
Physics-informed neural networks (PINNs), introduced in [M. Raissi, P. Perdikaris, and G. Karniadakis, J. Comput. Phys., 378 (2019), pp. 686--707], are effective in solving integer-order partial differential equations (PDEs) based on scattered and noisy data. PINNs employ standard feedforward neural networks (NNs) with the PDEs explicitly encoded into the NN using automatic differentiation, while the sum of the mean-squared PDE residuals and the mean-squared error in initial-boundary conditions is minimized with respect to the NN parameters. Here we extend PINNs to fractional PINNs (fPINNs) to solve space-time fractional advection-diffusion equations (fractional ADEs), and we study systematically their convergence, hence explaining both fPINNs and PINNs for the first time. Specifically, we demonstrate their accuracy and effectiveness in solving multidimensional forward and inverse problems with forcing terms whose values are only known at randomly scattered spatio-temporal coordinates ( black-box (BB) forcing terms). A novel element of the fPINNs is the hybrid approach that we introduce for constructing the residual in the loss function using both automatic differentiation for the integer-order operators and numerical discretization for the fractional operators. This approach bypasses the difficulties stemming from the fact that automatic differentiation is not applicable to fractional operators because the standard chain rule in integer calculus is not valid in fractional calculus. To discretize the fractional operators, we employ the Grünwald--Letnikov (GL) formula in one-dimensional fractional ADEs and the vector GL formula in conjunction with the directional fractional Laplacian in two- and three-dimensional fractional ADEs. We first consider the one-dimensional fractional Poisson equation and compare the convergence of the fPINNs against the finite difference method (FDM). We present the solution convergence using both the mean $L^2$ error as well as the standard deviation due to sensitivity to NN parameter initializations. Using different GL formulas we observe first-, second-, and third-order convergence rates for small size training sets but the error saturates for larger training sets. We explain these results by analyzing the four sources of numerical errors due to discretization, sampling, NN approximation, and optimization. The total error decays monotonically (below $10^{-5}$ for a third-order GL formula) but it saturates beyond that point due to the optimization error. We also analyze the relative balance between discretization and sampling errors and observe that the sampling size and the number of discretization points ( auxiliary points) should be comparable to achieve the highest accuracy. As we increase the depth of the NN up to certain value, the mean error decreases and the standard deviation increases, whereas the width has essentially no effect unless its value is either too small or too large. We next consider time-dependent fractional ADEs and compare white-box (WB) and BB forcing. We observe that for the WB forcing, our results are similar to the aforementioned cases; however, for the BB forcing fPINNs outperform FDM. Subsequently, we consider multidimensional time-, space-, and space-time-fractional ADEs using the directional fractional Laplacian and we observe relative errors of $10^{-3}\sim10^{-4}$. Finally, we solve several inverse problems in one, two, and three dimensions to identify the fractional orders, diffusion coefficients, and transport velocities and obtain accurate results given proper initializations even in the presence of significant noise.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“We also used the extended stochastic gradient descent Adam algorithm [18] to optimize the loss function.”
MethodRaissi, M., Perdikaris, P., & Karniadakis, G. E. (2019) · Journal of Computational Physics · doi:10.1016/j.jcp.2018.10.045 · Cited by 20,384 · cited 4×
“We just need to add the parameters to be identified in the inverse problem to the list of parameters to be optimized in the forward problem without changing anything else [35].”
BackgroundMethodInfluentialSelf-citationModel-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
Finn, C., Abbeel, P., & Levine, S. (2017) · arXiv · doi:10.48550/arxiv.1703.03400 · Cited by 8,512 · cited 1×
“We can tune these attributes of NN architecture automatically by leveraging metalearning techniques [65, 66].”
MethodImproving neural networks by preventing co-adaptation of feature detectors
Hinton, G. E., Srivastava, N., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. R. (2012) · arXiv · doi:10.48550/arxiv.1207.0580 · Cited by 6,382 · cited 1×
“Additionally, the noise in the data can be naturally taken into account by employing regularization techniques, such as L(1), L(2) and L∞ regularization [32], early stopping [33], as well as dropout [34, 35].”
MethodDiscovering governing equations from data by sparse identification of nonlinear dynamical systems
Brunton, S. L., Proctor, J. L., & Kutz, J. N. (2016) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1517384113 · Cited by 4,796 · cited 10×
“In this paper, we focus on the identification of the parameters of fractional PDEs whose overall form is known but some coefficients and most importantly the operators themselves are not known.”
BackgroundInfluentialNeural Architecture Search with Reinforcement Learning
Zoph, B., & Le, Q. V. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1611.01578 · Cited by 3,974 · cited 1×
“We can tune these attributes of NN architecture automatically by leveraging meta-learning techniques [54].”
MethodArtificial neural networks for solving ordinary and partial differential equations
Lagaris, I. E., Likas, A., & Fotiadis, D. I. (1998) · IEEE Transactions on Neural Networks · doi:10.1109/72.712178 · Cited by 2,412 · cited 1×
“In particular, we concentrate on physics-informed neural networks (PINNs) [7, 46, 19, 35], which belong to the second aforementioned category.”
BackgroundHidden physics models: Machine learning of nonlinear partial differential equations
Raissi, M., & Karniadakis, G. E. (2018) · Journal of Computational Physics · doi:10.1016/j.jcp.2017.11.039 · Cited by 1,349 · cited 1×
“Other examples include, but not limited to, [21, 22, 23].”
BackgroundSelf-citationApplication of a fractional advection‐dispersion equation
Benson, D. A., Wheatcraft, S. W., & Meerschaert, M. M. (2000) · Water Resources Research · doi:10.1029/2000wr900031 · Cited by 1,301 · cited 1×
“Among these are solute transport in fractured and porous media [2, 3], acoustic wave propagation with frequency-dependent dissipation [4, 5, 6, 7], laminar and turbulent flows [8, 9, 10], viscoelastic constitutive laws [11], just to mention a few.”
BackgroundA fully discrete difference scheme for a diffusion-wave system
Sun, Z.-Z., & Wu, X. (2006) · Applied Numerical Mathematics · doi:10.1016/j.apnum.2005.03.003 · Cited by 1,235 · cited 2×
“The truncating error for the L1 scheme is (\Delta t) 2 - \gamma [44].”
Method
Research works printed but not linked: 31
- [62] Avrim Blum and Ronald L Rivest. Training a 3-node neural network is np-complete. In Advances in neural information processing systems , pages 494-501, 1989. -
- [13] Samuel H Rudy, Steven L Brunton, Joshua L Proctor, and J Nathan Kutz. Data-driven discovery of partial differential equations. Science Advances , 3(4):e1602614, 2017. -
- [14] Zichao Long, Yiping Lu, Xianzhong Ma, and Bin Dong. PDE-net: Learning PDEs from data. In Jennifer Dy and Andreas Krause, editors, Proceedings of the 35th International Conference on Machine Learning , volume 80 of Proceedings of Machine Learning Research , pages 32083216, Stockholmsmässan, Stockholm Sweden, 10-15 Jul 2018. PMLR. -
- [11] Francesco Mainardi. Fractional calculus and waves in linear viscoelasticity: an introduction to mathematical models . World Scientific, 2010. -
- [5] Bradley E Treeby and BT Cox. Modeling power law absorption and dispersion for acoustic propagation using the fractional laplacian. The Journal of the Acoustical Society of America , 127(5):2741-2748, 2010. -
- 10.1007/978-3-642-14574-2 · doi:10.1007/978-3-642-14574-2
- [57] Rudolf Gorenflo, Anatoli˘ ı Aleksandrovich Kilbas, Francesco Mainardi, Sergei V Rogosin, et al. Mittag-Leffler functions, related topics and applications , volume 2. Springer, 2014. -
- [19] Vladimir I Gorbachenko, Tatiana V Lazovskaya, Dmitriy A Tarkhov, Alexander N Vasilyev, and Maxim V Zhukov. Neural network technique in some inverse problems of mathematical physics. In International Symposium on Neural Networks , pages 310-316. Springer, 2016. -
- [59] Xavier Glorot and Yoshua Bengio. Understanding the difficulty of training deep feedforward neural networks. In Proceedings of the thirteenth international conference on artificial intelligence and statistics , pages 249-256, 2010. -
- 10.1007/978-3-319-40663-3_36 · doi:10.1007/978-3-319-40663-3_36
- 10.1007/978-3-662-43930-2 · doi:10.1007/978-3-662-43930-2
- [47] A Anatolii Aleksandrovich Kilbas, Hari Mohan Srivastava, and Juan J Trujillo. Theory and applications of fractional differential equations , volume 204. Elsevier Science Limited, 2006. -
- [23] Mamikon Gulian, Maziar Raissi, Paris Perdikaris, and George Karniadakis. Machine learning of space-fractional differential equations. arXiv preprint arXiv:1808.00931 , 2018. - · arXiv
- [24] Tianping Chen and Hong Chen. Approximations of continuous functionals by neural networks with application to dynamic systems. IEEE Transactions on Neural Networks , 4(6):910-918, 1993. -
- [7] Tieyuan Zhu and Jerry M Harris. Modeling acoustic wave propagation in heterogeneous attenuating media using decoupled fractional laplacians. Geophysics , 79(3):T105-T116, 2014. -
- [54] Bangti Jin, Raytcho Lazarov, and Zhi Zhou. An analysis of the l1 scheme for the subdiffusion equation with nonsmooth data. IMA Journal of Numerical Analysis , 36(1):197-221, 2015. -
- Kilbas A. A. A., 2006, North-Holland Math. Stud. 204
- [61] Jason D Lee, Max Simchowitz, Michael I Jordan, and Benjamin Recht. Gradient descent only converges to minimizers. In Conference on Learning Theory , pages 1246-1257, 2016. -
- [67] Lihua Lei, Cheng Ju, Jianbo Chen, and Michael I Jordan. Non-convex finite-sum optimization via scsg methods. In Advances in Neural Information Processing Systems , pages 2348-2358, 2017. -
- [68] Chenxin Ma, Jakub Konečn` y, Martin Jaggi, Virginia Smith, Michael I Jordan, Peter Richtárik, and Martin Takáč. Distributed optimization with arbitrary local solvers. Optimization Methods and Software , 32(4):813-848, 2017.
- 10.1142/p614 · doi:10.1142/p614
- Meerschaert M. M., 2004, Fract. Calc. Appl. Anal., 7, 61
- [41] Ying Zhang and Xiang Xu. Inverse source problem for a fractional diffusion equation. Inverse problems , 27(3):035010, 2011. -
- [42] Bangti Jin and William Rundell. An inverse problem for a one-dimensional time-fractional diffusion problem. Inverse Problems , 28(7):075010, 2012. -
- [50] Igor Podlubny. Fractional differential equations: an introduction to fractional derivatives, fractional differential equations, to methods of their solution and some of their applications. Mathematics in science and engineering , 198:261-300, 1999. -
- [51] Kai Diethelm. The analysis of fractional differential equations: An application-oriented exposition using differential operators of Caputo type . Springer Science & Business Media, 2010. -
- [32] Yoshua Bengio, Ian J Goodfellow, and Aaron Courville. Deep learning. Nature , 521(7553):436444, 2015. -
- Samko S. G., 1993, Gordon and Breach
- [64] John Paul Roop. Computational aspects of fem approximation of fractional advection dispersion equations on bounded domains in r2. Journal of Computational and Applied Mathematics , 193(1):243-268, 2006. -
- [9] Vladimir V Uchaikin. Fractional derivatives for physicists and engineers , volume 2. Springer, 2013. -
- [35] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research , 15(1):1929-1958, 2014. -
Cited by
Items 101–200 of 1,000
Tanyu, D. N., Ning, J., Freudenberg, T., Heilenkötter, N., Rademacher, A., Iben, U., & Maaß, P. (2023) · Inverse Problems · doi:10.1088/1361-6420/ace9d4 · Cited by 112
“In addition, PINN can be extended to solve integro-differential equations [115], fractional PDEs [135] and stochastic PDEs [171].”
BackgroundZhang, Z., Zhang, H., Zhang, L., & Guo, L. (2023) · Journal of Computational Physics · doi:10.1016/j.jcp.2023.112415 · Cited by 90
“More targeted neural networks such as the Bayesian PINN [16], discrete PINN framework based on graph convolutional network and variational structure of PDE [17], extended physics-informed neural networks (XPINNs) [18], parareal physics-informed neural network (PPINN) [19], auxiliary physics informed neural networks (A…”
MethodDarbon, J., & Meng, T. (2020) · Journal of Computational Physics · doi:10.1016/j.jcp.2020.109907 · Cited by 85
Convergence Rate Analysis for Deep Ritz Method
Duan, C., Jiao, Y., Lai, Y., Li, D., Lu, X., & Yang, J. Z. (2022) · Communications in Computational Physics · doi:10.4208/cicp.oa-2021-0195 · Cited by 93
“[37, 27, 38] study the convergence of PINNs with deep neural networks.”
BackgroundInfluentialApplications of physics informed neural operators
Rosofsky, S., Majed, H. A., & Huerta, E. A. (2023) · Machine Learning Science and Technology · doi:10.1088/2632-2153/acd168 · Cited by 114
“Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics 378, 686 (2019) [28] G.”
BackgroundXu, R., Zhang, D., Miao, R., & Wang, N. (2021) · Journal of Computational Physics · doi:10.1016/j.jcp.2021.110318 · Cited by 73
“Later, variations of PINN have been used for a variety of forward and inverse problems, which require the solution of PDEs (Jagtap et al., 2020; Pang et al., 2018).”
Mathematical Introduction to Deep Learning: Methods, Implementations, and TheoryPreprint
Jentzen, A., Kuckuck, B., & von Wurstemberger, P. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2310.20360 · Cited by 83
Greedy training algorithms for neural networks and applications to PDEs
Siegel, J. W., Hong, Q., Jin, X., Hao, W., & Xu, J. (2023) · Journal of Computational Physics · doi:10.1016/j.jcp.2023.112084 · Cited by 71
“This approach has been applied to a wide variety of practical problems with astounding success [9, 10, 56, 70, 62].”
BackgroundPre-training strategy for solving evolution equations based on physics-informed neural networks
Guo, J., Yao, Y., Wang, H., & Gu, T.-X. (2023) · Journal of Computational Physics · doi:10.1016/j.jcp.2023.112258 · Cited by 71
“Pang et al. extended PINN to fractional PINN (fPINN) for solving fractional differential equations (FDEs) and systematically demonstrated the convergence and efficiency of the algorithm [14].”
BackgroundWang, J., Liu, Y., Yan, L., Han, K., Feng, L., & Zhang, R. (2025) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0259937 · Cited by 71
Wang, N., Chang, H., & Zhang, D. (2022) · Journal of Computational Physics · doi:10.1016/j.jcp.2022.111419 · Cited by 77
A Rate of Convergence of Physics Informed Neural Networks for the Linear Second Order Elliptic PDEs
Jiao, Y., Lai, Y., Li, D., Lu, X., Wang, F. W. F., Wang, Y., & Yang, J. Z. (2022) · Communications in Computational Physics · doi:10.4208/cicp.oa-2021-0186 · Cited by 88
“Both DRM and DGM are applied to variational forms of PDEs, and PINNs are based on residual minimization to the differential equation, see [2, 29, 17, 25], which can be extended to general PDEs [14, 16, 23, 22].”
Wu, G.-Z., Fang, Y., Kudryashov, N. A., Wang, Y., & Dai, C. (2022) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2022.112143 · Cited by 78
Wu, W., Daneker, M., Jolley, M. A., Turner, K. T., & Lu, L. (2023) · Applied Mathematics and Mechanics · doi:10.1007/s10483-023-2995-8 · Cited by 68
“Effective data sampling strategies and boundary condition constraints of PINNs 1041 PINNs have succeeded in estimating the solutions to a wide range of forward and inverse problems, including classic differential problems, fractional equations([34]), integral-differential equations([25,35]), and stochastic PDEs([36]).”
BackgroundSelf-citationPhysics-informed neural networks method in high-dimensional integrable systems
Miao, Z., & Chen, Y. (2021) · Modern Physics Letters B · doi:10.1142/s021798492150531x · Cited by 71
“The PINN method has even been applied to fractional partial differential equations [21], stochastic partial differential equations [22], and so on.”
MethodOommen, V., & Srinivasan, B. (2022) · Journal of Computing and Information Science in Engineering · doi:10.1115/1.4053800 · Cited by 66
Recipes for when physics fails: recovering robust learning of physics informed neural networks
Bajaj, C., McLennan, L., Andeen, T., & Roy, A. (2023) · Machine Learning Science and Technology · doi:10.1088/2632-2153/acb416 · Cited by 92
“Many variants of the trainable PINN architecture have been explored to exploit structure, quality and speed of convergence, and dimensional scalability [17, 18, 19, 20, 21, 22, 23].”
BackgroundChen, X., Duan, J., & Karniadakis, G. E. (2020) · European Journal of Applied Mathematics · doi:10.1017/s0956792520000169 · Cited by 75
“PINNs: Physics-informed neural networks for deterministic PDEs First, we briefly review the type of deep neural networks (DNNs) to solve deterministic partial differential equations (PDEs) and the corresponding inverse problem [1, 16].”
Self-citationScalable algorithms for physics-informed neural and graph networks
Shukla, K., Xu, M., Trask, N., & Karniadakis, G. E. (2022) · Data-Centric Engineering · doi:10.1017/dce.2022.24 · Cited by 76
“They can easily be applied to any known PDE, but also to different types of differential equations, including fractional PDEs (Pang et al., 2019), integro-differential equations (Lu et al.”
BackgroundInfluentialSelf-citationDeep learning phase‐field model for brittle fractures
Motlagh, Y. G., Jimack, P. K., & de Borst, R. (2022) · International Journal for Numerical Methods in Engineering · doi:10.1002/nme.7135 · Cited by 72
“(S ∶ BC) ⎧⎨⎪⎩ u = gi, x ∈ ∂Ωgi , σijnj = hi, x ∈ ∂Ωhi , ∂c ∂xi ni = 0, x ∈ ∂Ω, (22)”
BackgroundYang, X., Zhou, Z., Li, L., & Zhang, X. (2023) · Mechanism and Machine Theory · doi:10.1016/j.mechmachtheory.2023.105439 · Cited by 68
A data-driven physics-informed neural network for predicting the viscosity of nanofluids
Esfahani, I. C. (2023) · AIP Advances · doi:10.1063/5.0132846 · Cited by 65
Machine Learning for Clouds and Climate
Beucler, T., Ebert‐Uphoff, I., Rasp, S., Pritchard, M. S., & Gentine, P. (2023) · Geophysical monograph · doi:10.1002/9781119700357.ch16 · Cited by 67
Application of physics-informed neural networks for nonlinear buckling analysis of beams
Bazmara, M., Mianroodi, M., & Silani, M. (2023) · Acta Mechanica Sinica · doi:10.1007/s10409-023-22438-x · Cited by 64
Li, X., Deng, J., Wu, J., Zhang, S., Li, W., & Wang, Y. (2024) · Computers & Mathematics with Applications · doi:10.1016/j.camwa.2024.01.021 · Cited by 72
Physics-Informed Deep Neural Network for Inhomogeneous Magnetized Plasma Parameter Inversion
Zhang, Y., Fu, H., Qin, Y., Wang, K., & Ma, J. (2022) · IEEE Antennas and Wireless Propagation Letters · doi:10.1109/lawp.2022.3149889 · Cited by 60
Lejarza, F., & Bâldea, M. (2022) · Scientific Reports · doi:10.1038/s41598-022-13644-w · Cited by 71
A Taxonomic Survey of Physics-Informed Machine Learning
Ghosh, P., Pateras, J., & Rana, P. (2023) · Applied Sciences · doi:10.3390/app13126892 · Cited by 68
“Additionally, PINNs have been adapted to serve fractional expressions of differential equations with a Caputo–Hadamard augmentation [5,42].”
MethodA Physics-Informed Neural Network Framework for PDEs on 3D Surfaces: Time Independent Problems
Fang, Z., & Zhan, J. (2019) · IEEE Access · doi:10.1109/access.2019.2963390 · Cited by 60
“Numerical methods for fractional order PDEs has been studied in most recent work [23], [24].”
MethodWang, L., & Yan, Z. (2021) · Physica D Nonlinear Phenomena · doi:10.1016/j.physd.2021.133037 · Cited by 61
“The physics-informed neural network (PINN) approach [21, 22] were presented, and has been applied to many linear and nonlinear PDEs [23–27], as well as other types of equations such as the fractional PDEs [28], and stochastic differential equations [29].”
MethodInfluentialPhysics-informed neural networks for nonlinear bending of 3D functionally graded beam
Bazmara, M., Silani, M., Mianroodi, M., & sheibanian, M. (2023) · Structures · doi:10.1016/j.istruc.2023.01.115 · Cited by 73
Prediction of fluid flow in porous media by sparse observations and physics-informed PointNet
Kashefi, A., & Mukerji, T. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.08.006 · Cited by 59
“solving forward and inverse problems and its latter versions with various enhancements and extensions have been introduced such as fPINN [Pang et al., 2019], nPINN [Pang et al.”
MethodLearning Specialized Activation Functions for Physics-Informed Neural Networks
Wang, H. W. H., Lu, L., Song, S., & Huang, G. M. (2023) · Communications in Computational Physics · doi:10.4208/cicp.oa-2023-0058 · Cited by 72
“Physics-informed neural networks have emerged as a promising method for solving forward and inverse problems of integer-order PDEs [1, 2, 4, 5], fractional PDEs [49], and stochastic PDEs [50].”
MethodSelf-citationTowards a Unified theory of Fractional and Nonlocal Vector Calculus
D’Elia, M., Gulian, M., Olson, H., & Karniadakis, G. E. (2021) · Fractional Calculus and Applied Analysis · doi:10.1515/fca-2021-0057 · Cited by 58
Self-citationWang, S., Zhang, H., & Jiang, X. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.03.032 · Cited by 57
Mesh-Free Surrogate Models for Structural Mechanic FEM Simulation: A Comparative Study of Approaches
Hoffer, J. G., Geiger, B. C., Ofner, P., & Kern, R. (2021) · Applied Sciences · doi:10.3390/app11209411 · Cited by 61
Zhou, T., Zhang, X., Droguett, E. L., & Mosleh, A. (2022) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2022.108835 · Cited by 62
Alkhadhr, S., & Almekkawy, M. K. (2023) · Sensors · doi:10.3390/s23052792 · Cited by 67
“This makes finding the global minima an NP-hard problem [58].”
BackgroundShang, Y., Wang, F., & Sun, J. (2023) · Communications in Nonlinear Science and Numerical Simulation · doi:10.1016/j.cnsns.2023.107518 · Cited by 80
Physics-Informed Neural Networks in Materials Modeling and Design: A Review
Michaloglou, A., Papadimitriou, I., Gialampoukidis, I., Vrochidis, S., & Kompatsiaris, I. Y. (2025) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-025-10448-9 · Cited by 92
Zhou, Z., & Yan, Z. (2021) · Communications in Theoretical Physics · doi:10.1088/1572-9494/ac1cd9 · Cited by 59
Zhong, L., Wu, B., & Wang, Y. (2022) · Physics of Fluids · doi:10.1063/5.0106506 · Cited by 55
“Compared with FEM and FVM, PINNs are a mesh-free method which has high flexibility and can easily adapt to complex-geometry domain [14].”
BackgroundMethodSun, J., Dong, S. S., & Wang, F. (2024) · Journal of Computational and Applied Mathematics · doi:10.1016/j.cam.2024.115830 · Cited by 105
Mehta, P. P., Pang, G., Song, F., & Karniadakis, G. E. (2019) · Fractional Calculus and Applied Analysis · doi:10.1515/fca-2019-0086 · Cited by 55
Self-citationBilevel Parameter Learning for Nonlocal Image Denoising Models
D’Elia, M., Reyes, J. C. D. L., & Miniguano–Trujillo, A. (2021) · Journal of Mathematical Imaging and Vision · doi:10.1007/s10851-021-01026-2 · Cited by 55
“local kernel identification, now subject of great interest in the nonlocal community, and provides an alternative to neural-networks-based algorithms [31].”
BackgroundAdaptive deep density approximation for Fokker-Planck equations
Tang, K., Wan, X., & Liao, Q. (2022) · Journal of Computational Physics · doi:10.1016/j.jcp.2022.111080 · Cited by 77
“In [10, 11, 12, 13], physics-informed neural networks are developed through infusing PDEs into networks as a constraint.”
Sun, T., Yao, C., Liu, Z., Huang, S., Huang, X., Zheng, S., Liu, J., Shi, P., Zhang, T., Chen, H., et al. (2024) · Nano Energy · doi:10.1016/j.nanoen.2024.109395 · Cited by 56
Stiff-PDEs and Physics-Informed Neural Networks
Sharma, P., Evans, L. M., Tindall, M., & Nithiarasu, P. (2023) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-023-09890-4 · Cited by 110
“Other than baseline PINNs, DeepXDE can also solve forward and inverse integro-differential equations, (IDEs) [92], fractional PDEs, (fPDEs) [92], stochastic PDEs, (sPDEs) [93], topology optimization with hard constraints, (hPINN) [7], PINN with multi-scale Fourier features [94] and multifidelity NN, (MFNN) [95, 96].”
BackgroundSivalingam, S. M., Kumar, P., & Govindaraj, V. (2023) · Mathematics and Computers in Simulation · doi:10.1016/j.matcom.2023.06.012 · Cited by 53
Degen, D., Caviedes‐Voullième, D., Buiter, S., Franssen, H. H., Vereecken, H., González‐Nicolás, A., & Wellmann, F. (2023) · Geoscientific model development · doi:10.5194/gmd-16-7375-2023 · Cited by 65
“…Bézenac et al., 2019; Dwivedi and Srini-vasan, 2020; Geneva and Zabaras, 2020; Karumuri et al., 2020; Meng and Karniadakis, 2020; Meng et al., 2020; Pang et al., 2019, 2020; Peng et al., 2020; Raissi et al., 2019; Shah et al., 2019; Sharma et al., 2018; Wu et al., 2020; Yang et al., 2018; Yang…”
BackgroundInfluentialChen, Z., Lai, S.-K., Yang, Z., Ni, Y., Yang, Z., & Cheung, K. C. (2025) · Computer Methods in Applied Mechanics and Engineering · doi:10.1016/j.cma.2024.117691 · Cited by 53
MYCRUNCHGPT: A LLM ASSISTED FRAMEWORK FOR SCIENTIFIC MACHINE LEARNING
Kumar, V., Gleyzer, L., Kahana, A., Shukla, K., & Karniadakis, G. E. (2023) · Journal of Machine Learning for Modeling and Computing · doi:10.1615/jmachlearnmodelcomput.2023049518 · Cited by 79
“Since their inception, many variants of PINNs have been developed, including (but not limited to) conservative, Bayesian [32], fractional [33], parareal [31], variational [34], stochastic [35], among others.”
BackgroundSelf-citationQiu, L., Wang, Y., Gu, Y., Qin, Q.-H., & Wang, F. (2024) · Applied Mathematical Modelling · doi:10.1016/j.apm.2024.115906 · Cited by 52
Ye, Y., Fan, H., Li, Y., Liu, X., & Zhang, H. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.08.030 · Cited by 51
Fractional physics-informed neural networks for time-fractional phase field models
Wang, S., Zhang, H., & Jiang, X. (2022) · Nonlinear Dynamics · doi:10.1007/s11071-022-07746-3 · Cited by 59
“Bandai & Ghezzehei [54] employed PINNs to solve the inverse problem of the Richardson– Richards equation without initial and boundary conditions, and fPINNs to obtain the order of fractional derivative and diffusion coefficient of the space–time fractional advection–diffusion equation from observations was derived in…”
BackgroundMethodZhong, M., Gong, S., Tian, S., & Yan, Z. (2022) · Physica D Nonlinear Phenomena · doi:10.1016/j.physd.2022.133430 · Cited by 51
Deep learning for thermal plasma simulation: Solving 1-D arc model as an example
Zhong, L., Gu, Q., & Wu, B. (2020) · Computer Physics Communications · doi:10.1016/j.cpc.2020.107496 · Cited by 49
Physics Informed Neural Networks (PINNs) For Approximating Nonlinear Dispersive PDEs
Bai, G., Koley, U., Mishra, S., & Molinaro, R. (2021) · Journal of Computational Mathematics · doi:10.4208/jcm.2101-m2020-0342 · Cited by 49
Contaminant source identification in groundwater by means of artificial neural network
Secci, D., Molino, L., & Zanini, A. (2022) · Journal of Hydrology · doi:10.1016/j.jhydrol.2022.128003 · Cited by 48
“This modeling technique already has achieved good results in the literature (Pang et al., 2019; Mao et al., 2020; Wang et al., 2021) and this motivates the possibility of using it for transport forward and inverse problems which are described by partial differential equations.”
MethodUsama, M., Ma, R., Hart, J., & Wojcik, M. (2022) · Algorithms · doi:10.3390/a15120447 · Cited by 48
Physics-Informed neural network solver for numerical analysis in geoengineering
Chen, X., Zhang, P., & Yin, Z. (2024) · Georisk Assessment and Management of Risk for Engineered Systems and Geohazards · doi:10.1080/17499518.2024.2315301 · Cited by 68
“Regarding the randomly distributed data, various sampling methods including normal sampling, Latin hypercube sampling (Rao, Sun, and Liu 2021), Sobol ’ sequence (Pang, Lu, and Karniadakis 2019; Guo et al. 2022), Halton sequence and Hammers-ley sequence (Wu et al. 2023) have been used.”
MethodA Gaussian mixture distribution-based adaptive sampling method for physics-informed neural networks
Jiao, Y., Li, D., Lu, X., Yang, J. Z., & Yuan, C. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.108770 · Cited by 52
“To visualize the adaptive training process in GAS, the loss and L 2 error with different N G are also depicted in Figure 14.”
MethodInfluentialA practitioner’s guide to Kolmogorov–Arnold networks
Noorizadegan, A., Wang, S., Ling, L., & Dominguez-Morales, J. P. (2026) · Computer Science Review · doi:10.1016/j.cosrev.2026.100991 · Cited by 55
Advancing Fusion with Machine Learning Research Needs Workshop Report
Humphreys, D., Kupresanin, A., Boyer, M. D., Canik, J., Chang, C. S., Cyr, E. C., Granetz, R., Hittinger, J., Kolemen, E., Lawrence, E., et al. (2020) · Journal of Fusion Energy · doi:10.1007/s10894-020-00258-1 · Cited by 47
D’Elia, M., Gunzburger, M., & Vollmann, C. (2021) · Mathematical Models and Methods in Applied Sciences · doi:10.1142/s0218202521500317 · Cited by 47
On quadrature rules for solving Partial Differential Equations using Neural Networks
Rivera, J. A., Taylor, J. M., Omella, Á. J., & Pardo, D. (2022) · Computer Methods in Applied Mechanics and Engineering · doi:10.1016/j.cma.2022.114710 · Cited by 56
“However, DL also presents limitations when solving PDEs.”
InfluentialOn physics-informed neural networks for quantum computers
Markidis, S. (2022) · Frontiers in Applied Mathematics and Statistics · doi:10.3389/fams.2022.1036711 · Cited by 61
“A similar approach is used in classical fractional PINN [47].”
MethodSurrogate modeling for Bayesian inverse problems based on physics-informed neural networks
Li, Y., Wang, Y., & Yan, L. (2022) · Journal of Computational Physics · doi:10.1016/j.jcp.2022.111841 · Cited by 45
A class of improved fractional physics informed neural networks
Ren, H., Meng, X., Liu, R., Hou, J., & Yu, Y. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126890 · Cited by 49
Ramirez, I., Gómez, J. P., Pardo, D., Sanz, M., del Rio, L., Ortiz, A., Morozovska, K., & Aizpurua, J. I. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109556 · Cited by 44
“That is, the complexity of real problems (Pang et al., 2019; Kapoor et al., 2023; Kang et al., 2022), and convergence issues (Xiang et al., 2022; Anagnostopoulos et al., 2024b), have been key limitations addressed in the PINN literature.”
Physics-Informed Neural Networks for PDE-Constrained Optimization and Control
Barry-Straume, J., Sarshar, A., Popov, A. A., & Sandu, A. (2025) · Communications on Applied Mathematics and Computation · doi:10.1007/s42967-025-00499-x · Cited by 44
“Since their inception PINNs have been applied to solve a wide range of problems including PDEs [19, 34, 47, 49], inverse problems [6, 15, 22, 32, 33, 35], solution of fractional differential equations [31], and stochastic differential equations [27, 48, 50, 51].”
PINN-based approach to the consolidation analysis of visco-elastic soft soil around twin tunnels
Xie, S., Hu, A., Xiao, Z., Mariani, S., & Vecchia, G. D. (2024) · Tunnelling and Underground Space Technology · doi:10.1016/j.tust.2024.105981 · Cited by 43
Pang, G., & Karniadakis, G. E. (2020) · Nonlinear systems and complexity · doi:10.1007/978-3-030-44992-6_14 · Cited by 54
Self-citationWu, Y., Aghamiry, H. S., Operto, S., & Ma, J. (2023) · Geophysics · doi:10.1190/geo2022-0479.1 · Cited by 42
Physics-informed neural networks for mesh deformation with exact boundary enforcement
Aygun, A., Maulik, R., & Karakus, A. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.106660 · Cited by 42
“Applications of PINN cover the solutions of conservation laws [22], fractional and stochastic differential equations [23, 24], solution of Navier-Stokes equations [25, 26], Euler equations [27], heat transfer problems [28, 29], Boltzmann equation with Bhatnagar-Gross-Krook collision model [30],…”
MethodZheng, H., Huang, Y. T., Huang, Z., Hao, W., & Lin, G. (2023) · Journal of Computational Physics · doi:10.1016/j.jcp.2023.112751 · Cited by 44
“Building upon PINNs, Pang et al. [10] developed fractional PINNs (fPINNs) for solving space-time fractional advection-di ff usion equations from scattered and noisy data.”
Taghizadeh, M., Nabian, M. A., & Alemazkoor, N. (2024) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.13312 · Cited by 42
“By incorporating a loss function that measures how well the output satisfies the equations (Raissi et al., 2019), PINNs can tackle both forward and inverse problems, extracting physical states from measurements (Mao et al., 2020; Nabian & Meidani, 2020; Pang et al., 2019).”
Xiang, Z., Peng, W., Yao, W., & Zhou, W. (2025) · Journal of Thermophysics and Heat Transfer · doi:10.2514/1.t7077 · Cited by 42
Di, D., Bai, Y., Fang, H., Sun, B., Wang, N., & Li, B. (2025) · Automation in Construction · doi:10.1016/j.autcon.2025.106246 · Cited by 42
One-shot learning for solution operators of partial differential equations
Jiao, A., He, H., Ranade, R., Pathak, J., & Lu, L. (2025) · Nature Communications · doi:10.1038/s41467-025-63076-z · Cited by 54
“For example, we can enforce physics-based constraints to train neural networks that learn the underlying physics (Pang et al., 2019; Zhang et al., 2019; Chen et al., 2020a; Yazdani et al., 2020; Rao et al., 2020; Wu et al., 2018; Qian et al., 2020; Lu et al., 2021b).”
Self-citationThermodynamics of Learning Physical Phenomena
Cueto, E., & Chinesta, F. (2023) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-023-09954-5 · Cited by 50
“Physics-informed neural networks (PINNs) have constituted a great success in computational mechanics and mathematics and are, perhaps, the responsible of a very active research activity in the field [30, 45, 46, 47, 48, 49, 50, 51, 52].”
BackgroundComplex dynamics on the one-dimensional quantum droplets via time piecewise PINNs
Pu, J., & Chen, Y. (2023) · Physica D Nonlinear Phenomena · doi:10.1016/j.physd.2023.133851 · Cited by 42
“Generally, we consider a multi-dimensional spatiotemporal real nonlinear system whose governing equations can be described by a set of coupled, parameterized and variable coefficient PDEs in the general form given by in which q = q ( x , t ) ∈ 1 n is the n -dimensional latent solution, t ∈ [0 , T ] denotes time and x ∈…”
MethodZhang, Z., Li, J., & Liu, B. (2024) · Journal of Computational Physics · doi:10.1016/j.jcp.2024.113561 · Cited by 41
Physics-Informed Neural Networks
Kollmannsberger, S., D’Angella, D., Jokeit, M., & Herrmann, L. (2021) · Studies in computational intelligence · doi:10.1007/978-3-030-76587-3_5 · Cited by 51
Wasserstein generative adversarial uncertainty quantification in physics-informed neural networks
Gao, Y., & Ng, M. K. (2022) · Journal of Computational Physics · doi:10.1016/j.jcp.2022.111270 · Cited by 50
“, fractional PDEs [43], stochastic PDEs [15, 55, 58] as well as high dimensional problems [23] etc.”
BackgroundThe Neural Network Approach to Inverse Problems in Differential EquationsPreprint
Xu, K., & Darve, E. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1901.07758 · Cited by 43
“Pioneer works in [34] inspire the latter view.”
BackgroundSong, J. M., & Yan, Z. (2023) · Physica D Nonlinear Phenomena · doi:10.1016/j.physd.2023.133729 · Cited by 40
“And the PINNs method has been extended to solve the stochastic differential equations, fractional differential equations, and integro-differential equations [40–43], and soliton equations [44–46].”
MethodZhao, N., Chen, Y., Cheng, L., & Chen, J. (2024) · Nonlinear Dynamics · doi:10.1007/s11071-024-10562-6 · Cited by 40
“Based on the feasibility and validity of this method, PINN have been extensively applied to various types of PDEs, including integro-differential equations [10], fractional PDEs [11], and stochastic PDEs [12,13].”
Zhang, T., Zhang, D., Shi, S., & Guo, Z. (2025) · Nonlinear Dynamics · doi:10.1007/s11071-025-11067-6 · Cited by 39
“To overcome this issue, Pang et al. [41] proposed fractional PINNs (fPINNs), which are a hybrid approach that discretizes the fractional-order operators by traditional numerical algorithm while handling the integer-order operators using automatic differentiation.”
InfluentialQuan, W., Ma, X., Shang, Z., Zhao, K., Su, M., Dong, Z., & Zhao, Z. (2025) · Construction and Building Materials · doi:10.1016/j.conbuildmat.2025.142179 · Cited by 39
Energy-Based Error Bound of Physics-Informed Neural Network Solutions in Elasticity
Guo, M., & Haghighat, E. (2022) · Journal of Engineering Mechanics · doi:10.1061/(asce)em.1943-7889.0002121 · Cited by 56
“Moreover, the PINNs have also been investigated in their variational or fractional forms [20, 21, 28, 38].”
BackgroundPINN training using biobjective optimization: The trade-off between data loss and residual loss
Heldmann, F., Berkhahn, S., Ehrhardt, M., & Klamroth, K. (2023) · Journal of Computational Physics · doi:10.1016/j.jcp.2023.112211 · Cited by 39
Physics-guided Bayesian neural networks by ABC-SS: Application to reinforced concrete columns
Fernández, J., Chiachío, J., Chiachío, M., Barros, J., & Corbetta, M. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105790 · Cited by 43
Song, J. M., Zhong, M., Karniadakis, G. E., & Yan, Z. (2024) · Journal of Computational Physics · doi:10.1016/j.jcp.2024.112917 · Cited by 42
Self-citationStatistical solution to SDEs with $$\alpha $$-stable Lévy noise via deep neural network
Zhang, H., Xu, Y., Li, Y., & Kurths, J. (2020) · International Journal of Dynamics and Control · doi:10.1007/s40435-020-00677-0 · Cited by 38
Learning aerodynamics with neural network
Peng, W., Zhang, Y., Laurendeau, É., & Desmarais, M. C. (2022) · Scientific Reports · doi:10.1038/s41598-022-10737-4 · Cited by 42
“With the physical constrains, the PINNs manage to achieve state-of-the-art prediction accuracy 6,17–21 .”
BackgroundAccelerating heat exchanger design by combining physics-informed deep learning and transfer learning
Wu, Z., Zhang, B., Yu, H., Ren, J., Pan, M. H., He, C., & Chen, Q. (2023) · Chemical Engineering Science · doi:10.1016/j.ces.2023.119285 · Cited by 42
Spectral integrated neural networks (SINNs) for solving forward and inverse dynamic problems
Qiu, L., Wang, F., Qu, W., Gu, Y., & Qin, Q. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.106756 · Cited by 37
Zhang, H., Jiang, L., Chu, X., Wen, Y., Li, L., Liu, J., Xiao, Y., & Wang, L. (2024) · Computer Physics Communications · doi:10.1016/j.cpc.2024.109462 · Cited by 40
Improved Deep Neural Networks with Domain Decomposition in Solving Partial Differential Equations
Wu, W., Feng, X., & Xu, H. (2022) · Journal of Scientific Computing · doi:10.1007/s10915-022-01980-y · Cited by 37
“Andthe FPINNs[24]basedonafractionalequationidea.”
Method
Metadata
- Authors
- Pang, G., Lu, L., & Karniadakis, G. E.
- Venue
- SIAM Journal on Scientific Computing
- Published
- 2019-01
- Volume
- 41
- Issue
- 4
- Pages
- A2603-A2626
- ISSN
- 1064-8275 (ISSN-L)
- 1095-7197
- OpenAlex
- W2900501930
- Cited by
- 1,105
- References
- 38
- Citations by year
- 2026: 253 · 2025: 260 · 2024: 195 · 2023: 155 · 2022: 119 · 2021: 82 · 2020: 35 · 2019: 6
- FWCI
- 28.01
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- SIAM Journal on Scientific Computing (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Army Research Office (W911NF-18-1-0301, W911NF-15-1-0562)
- U.S. Department of Energy (DE-SC0019434, DE-SC0019453)
- Topics
- Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Model Reduction and Neural Networks
- Physical Sciences › Mathematics › Modeling and Simulation › Fractional Differential Equations Solutions
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- Keywords
- physics-informed neural networks, fractional advection-diffusion equation, fractional Poisson equation, convergence analysis, inverse problems, sampling error, optimization error
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.