Predictive uncertainty quantification for Bayesian Physics-Informed Neural Network (PINN) in hypocentre estimation problem

Date

2022-01-01

Author

Izzatullah, M.
Yıldırım, İbrahim Soner
Waheed, U.B.
Alkhalifah, T.

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

143
views

0
downloads

Physics-informed neural networks (PINNs) have appeared on the scene as a flexible and a versatile framework for solving partial differential equations (PDEs), along with any initial or boundary conditions. An important component of these solutions, especially when using the data as a boundary condition, is our confidence in their accuracy. There has been little study of PINN accuracy as an inversion tool. We introduce an approximate Bayesian framework for estimating predictive uncertainties in Physics-Informed Neural Network (PINN). This work investigates propagation of uncertainties from the random realisations of Eikonal PINN's weights and biases using the Laplace approximation. Laplace approximation is arguably the simplest family of approximations for the intractable posteriors of deep neural networks. We specifically use a hypocenter estimation problem based on the eikonal equation to demonstrate the approach effectiveness in measuring the predictive uncertainty in the PINN hypocenter estimation. The uncertainties estimation from this approach is called predictive uncertainty or, simply, forward modelling uncertainty in the context of PINN.

URI

https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85142904777&origin=inward
https://hdl.handle.net/11511/101646

Conference Name

83rd EAGE Conference and Exhibition 2022

Collections

Department of Computer Education and Instructional Technology, Conference / Seminar

Suggestions

OpenMETU
Core

Least squares differential quadrature time integration scheme in the dual reciprocity boundary element method solution of convection-diffusion problems Bozkaya, Canan (2005-03-18) The least squares differential quadrature method (DQM) is used for solving the ordinary differential equations in time, obtained from the application of the dual reciprocity boundary element method (DRBEM) for the spatial partial derivatives in convection-diffusion type problems. The DRBEM enables us to use the fundamental solution of the Laplace equation which is easy to implement computationally. The time derivative and the convection terms are considered as the nonhomogeneity in the equation which are ap...
Backward stochastic differential equations and Feynman-Kac formula in the presence of jump processes İncegül Yücetürk, Cansu; Yolcu Okur, Yeliz; Hayfavi, Azize; Department of Financial Mathematics (2013) Backward Stochastic Differential Equations (BSDEs) appear as a new class of stochastic differential equations, with a given value at the terminal time T. The application area of the BSDEs is conceptually wide which is known only for forty years. In financial mathematics, El Karoui, Peng and Quenez have a fundamental and significant article called “Backward Stochastic Differential Equations in Finance” (1997) which is taken as a groundwork for this thesis. In this thesis we follow the following steps: Firstl...
Collaborative Direction of Arrival estimation by using Alternating Direction Method of Multipliers in distributed sensor array networks employing Sparse Bayesian Learning framework Nurbas, Ekin; Onat, Emrah; Tuncer, Temel Engin (2022-10-01) In this paper, we present a new method for Direction of Arrival (DoA) estimation in distributed sensor array networks by using Alternating Direction Method of Multipliers (ADMM) in Sparse Bayesian Learning (SBL) framework. Our proposed method, CDoAE, has certain advantages compared to previous distributed DoA estimation methods. It does not require any special array geometry and there is no need for inter -array frequency and phase matching. CDoAE uses the distributed ADMM to update the parameter set extrac...
Distance-based discretization of parametric signal manifolds Vural, Elif (2010-06-28) The characterization of signals and images in manifolds often lead to efficient dimensionality reduction algorithms based on manifold distance computation for analysis or classification tasks. We propose in this paper a method for the discretization of signal manifolds given in a parametric form. We present an iterative algorithm for the selection of samples on the manifold that permits to minimize the average error in the manifold distance computation. Experimental results with image appearance manifolds d...
Semi-Bayesian Inference of Time Series Chain Graphical Models in Biological Networks Farnoudkia, Hajar; Purutçuoğlu Gazi, Vilda (null; 2018-09-20) The construction of biological networks via time-course datasets can be performed both deterministic models such as ordinary differential equations and stochastic models such as diffusion approximation. Between these two branches, the former has wider application since more data can be available. In this study, we particularly deal with the probabilistic approaches for the steady-state or deterministic description of the biological systems when the systems are observed though time. Hence, we consider time s...

Citation Formats

M. Izzatullah, İ. S. Yıldırım, U. B. Waheed, and T. Alkhalifah, “Predictive uncertainty quantification for Bayesian Physics-Informed Neural Network (PINN) in hypocentre estimation problem,” Madrid, Virtual, İspanya, 2022, vol. 1, Accessed: 00, 2023. [Online]. Available: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85142904777&origin=inward.