Variational Autoencoding of PDE Inverse Problems

06/28/2020
by   Daniel J. Tait, et al.
0

Specifying a governing physical model in the presence of missing physics and recovering its parameters are two intertwined and fundamental problems in science. Modern machine learning allows one to circumvent these, via emulators and surrogates, but in doing so disregards prior knowledge and physical laws that are especially important for small data regimes, interpretability, and decision making. In this work we fold the mechanistic model into a flexible data-driven surrogate to arrive at a physically structured decoder network. This provides accelerated inference for the Bayesian inverse problem, and can act as a drop-in regulariser that encodes a-priori physical information. We employ the variational form of the PDE problem and introduce stochastic local approximations as a form of model based data augmentation. We demonstrate both the accuracy and increased computational efficiency of the framework on real world settings and structured spatial processes.

READ FULL TEXT
research
10/21/2014

Variational Reformulation of Bayesian Inverse Problems

The classical approach to inverse problems is based on the optimization ...
research
10/17/2022

Data-Driven Joint Inversions for PDE Models

The task of simultaneously reconstructing multiple physical coefficients...
research
12/09/2019

Hybrid Physical-Deep Learning Model for Astronomical Inverse Problems

We present a Bayesian machine learning architecture that combines a phys...
research
07/04/2023

Hybrid two-level MCMC for Bayesian Inverse Problems

We introduced a novel method to solve Bayesian inverse problems governed...
research
09/25/2019

Structured random sketching for PDE inverse problems

For an overdetermined system Ax≈b with A and b given, the least-square (...
research
10/12/2022

Self-Validated Physics-Embedding Network: A General Framework for Inverse Modelling

Physics-based inverse modeling techniques are typically restricted to pa...

Please sign up or login with your details

Forgot password? Click here to reset