Abstract
Today, Physics-informed machine learning (PIML) methods are one of the effective tools with high flexibility for solving inverse problems and operational equations. Among these methods, physics-informed learning model built upon Gaussian processes (PIGP) has a special place due to provide the posterior probabilistic distribution of their predictions in the context of Bayesian inference. In this method, the training phase to determine the optimal hyper parameters is equivalent to the optimization of a non-convex function called the likelihood function. Due to access the explicit form of the gradient, it is recommended to use conjugate gradient (CG) optimization algorithms. In addition, due to the necessity of computation of the determinant and inverse of the covariance matrix in each evaluation of the likelihood function, it is recommended to use CG methods in such a way that it can be completed in the minimum number of evaluations. In previous studies, only special form of CG method has been considered, which naturally will not have high efficiency. In this paper, the efficiency of the CG methods for optimization of the likelihood function in PIGP has been studied. The results of the numerical simulations show that the initial step length and search direction in CG methods have a significant effect on the number of evaluations of the likelihood function and consequently on the efficiency of the PIGP. Also, according to the specific characteristics of the objective function in this problem, in the traditional CG methods, normalizing the initial step length to avoid getting stuck in bad conditioned points and improving the search direction by using angle condition to guarantee global convergence have been proposed. The results of numerical simulations obtained from the investigation of seven different improved CG methods with different angles in angle condition (four angles) and different initial step lengths (three step lengths), show the significant effect of the proposed modifications in reducing the number of iterations and the number of evaluations in different types of CG methods. This increases the efficiency of the PIGP method significantly, especially when the traditional CG algorithms fail in the optimization process, the improved algorithms perform well. Finally, in order to make it possible to implement the studies carried out in this paper for other parametric equations, the compiled package including the methods used in this paper is attached.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Data availability
The source code that supports the findings of this study are available on GitHub at https://github.com/massezati/Classical-CG-Normalized-CG-Normalized-with-Restart-CG-Methods.git.
References
Liu X, Yao W, Peng W, Zhou W (2023) Bayesian physics-informed extreme learning machine for forward and inverse PDE problems with Noisy data. Neurocomputing 549:126425
Donnelly J, Daneshkhah A, Abolfathi S (2024) Forecasting global climate drivers using Gaussian processes and convolutional autoencoders. Eng Appl Artif Intell 128:107536
Dwivedi V, Srinivasan B (2020) Physics informed extreme learning machine (PIELM)-a rapid method for the numerical solution of partial differential equations. Neurocomputing 391:96–118
Xiang Z, Peng W, Liu X, Yao W (2022) Self-adaptive loss balanced Physics-informed neural networks. Neurocomputing 496:11–34
Escapil-Inchauspe P, Ruz GA (2023) Hyper-parameter tuning of physics-informed neural networks: application to Helmholtz problems. Neurocomputing 561:126826
Chatrabgoun O, Esmaeilbeigi M, Cheraghi M, Daneshkhah A (2022) Stable likelihood computation for machine learning of linear differential operators with Gaussian processes. Int J Uncertain Quantif 12(3):75–99
Donnelly J, Daneshkhah A, Abolfathi S (2024) Physics-informed neural networks as surrogate models of hydrodynamic simulators. Sci Total Environ 912:168814
Hao Z, Liu S, Zhang Y, Ying C, Feng Y, Su H, Zhu J (2022) Physics-informed machine learning: a survey on problems, methods and applications. arXiv preprint arXiv:2211.08064
Asrav T, Aydin E (2023) Physics-informed recurrent neural networks and hyper-parameter optimization for dynamic process systems. Comput Chem Eng 173:108195
Yang X, Tartakovsky G, Tartakovsky A (2018) Physics-informed kriging: a physics-informed Gaussian process regression method for data-model convergence. arXiv preprint arXiv:1809.03461
Alvarez MA, Luengo D, Lawrence ND (2013) Linear latent force models using Gaussian processes. IEEE Trans Pattern Anal Mach Intell 35(11):2693–2705
Raissi M, Perdikaris P, Em Karniadakis G (2017) Machine learning of linear differential equations using Gaussian processes. J Comput Phys 348(1):683–693
Williams CK, Rasmussen CE (2006) Gaussian processes for machine learning. MIT Press, Cambridge
Esmaeilbeigi M, Cheraghi M (2023) Hybrid kernel approach to improving the numerical stability of machine learning for parametric equations with Gaussian processes in the noisy and noise-free data assumptions. Eng Comput 1–34. https://doi.org/10.1007/s00366-023-01818-7
Narayan A, Yan L, Zhou T (2021) Optimal design for kernel interpolation: applications to uncertainty quantification. J Comput Phys 430:1–20
Qin T, Chen Z, Jakeman JD, Xiu D (2021) Deep learning of parameterized equations with applications to uncertanity quantification. Int J Uncertain Quantif 11(2):63–82
Nocedal J, Wright SJ (eds) (1999) Numerical optimization. Springer, New York
Fletcher R, Reeves CM (1964) Function minimization by conjugate gradients. Comput J 7(2):149–154
Polak E, Ribiere G (1969) Note sur la convergence de methodes de directions conjuguees. Revue francaise dinformatique et de recherche operationnelle. Serie Rouge 3(16):35–43
Hestenes MR, Stiefel E (1952) Methods of conjugate gradients for solving linear systems. J Res Natl Bur Stand 49(6):409–436
Fletcher R (2000) Practical methods of optimization. Wiley, New York
Liu Y, Storey C (1991) Efficient generalized conjugate gradient algorithms, part 1: theory. J Optim Theory Appl 69:129–137
Dai YH, Yuan Y (1999) A nonlinear conjugate gradient method with a strong global convergence property. SIAM J Optim 10(1):177–182
Hager WW, Zhang H (2005) A new conjugate gradient method with guaranteed descent and an efficient line search. SIAM J Optim 16(1):170–192
Barzilai J, Borwein JM (1988) Two-point step size gradient methods. IMA J Numer Anal 8(1):141–148
Hager WW, Zhang H (2006) A survey of nonlinear conjugate gradient methods. Pac J Optim 2(1):35–58
Raissi M, Perdikaris P, Em Karniadakis G (2018) Numerical Gaussian processes for time-dependent and nonlinear partial differential equations. SIAM J Sci Comput 40(1):172–198
Perkins TJ, Jaeger J, Reinitz J, Glass L (2006) Reverse engineering the gap gene network of Drosophila melanogaster. PLoS Comput Biol 2(5):e51
Poustelnikova E, Pisarev A, Blagov M, Samsonova M, Reinitz J (2004) A database for management of gene expression data in situ. Bioinformatics 20(14):2212–2221
Acknowledgements
The authors would like to forward their sincere thanks to editor and anonymous referees, who spend their precious time in reviewing this paper.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no any Conflict of interest.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Ezati, M., Esmaeilbeigi, M. & Kamandi, A. Novel approaches for hyper-parameter tuning of physics-informed Gaussian processes: application to parametric PDEs. Engineering with Computers 40, 3175–3194 (2024). https://doi.org/10.1007/s00366-024-01970-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00366-024-01970-8