Ensemble Kalman Inversion

Ensemble Kalman Inversion (EKI) is a powerful Bayesian inference method used to solve inverse problems, particularly those involving complex, high-dimensional systems. Current research focuses on applying EKI in conjunction with various machine learning models, such as Physics-Informed Neural Operators (PINOs) and DeepONets, to improve efficiency and accuracy in diverse fields like reservoir characterization and system identification. This approach offers significant advantages in uncertainty quantification and computational speed compared to traditional methods, enabling faster and more reliable solutions for challenging inverse problems across scientific and engineering disciplines. The combination of EKI with generative models further enhances the ability to handle non-Gaussian posterior distributions and improve the quality of reconstructions.

Papers

June 2, 2024

Reservoir History Matching of the Norne field with generative exotic priors and a coupled Mixture of Experts -- Physics Informed Neural Operator Forward Model
Clement Etienam, Yang Juntao, Oleg Ovcharenko, Issam Said
Uncertainty Quantification Limited Field Generative Prior Physic Informed Neural Operator Reservoir Pressure Reservoir Model Critical Mixture Ratio Ensemble Kalman Inversion

April 20, 2024

A Novel A.I Enhanced Reservoir Characterization with a Combined Mixture of Experts -- NVIDIA Modulus based Physics Informed Neural Operator Forward Model
Clement Etienam, Yang Juntao, Issam Said, Oleg Ovcharenko, Kaustubh Tangsali, Pavel Dimitrov, Ken Hester
Mixture Component Expert Knowledge Reservoir Dynamic Physic Informed Neural Operator Reservoir Pressure Reservoir Model Ensemble Kalman Inversion State of the Art NVIDIA

March 6, 2024

Uncertainty quantification for deeponets with ensemble kalman inversion
Andrew Pensoneault, Xueyu Zhu
Uncertainty Quantification Operator Learning VI DeepONet Ensemble Kalman Inversion

October 2, 2023

Subsurface Characterization using Ensemble-based Approaches with Deep Generative Models
Jichao Bao, Hongkyu Yoon, Jonghyun Lee
Ensemble Based Subsurface Property Ensemble Kalman Inversion Regularized Newton Raphson Inversion

July 15, 2023

Gradient-free training of neural ODEs for system identification and control using ensemble Kalman inversion
Lucas Böttcher
External Control Gradient Based Neural ODE Neural Ordinary Differential Equation System Identification Gradient Free Ensemble Kalman Inversion

March 13, 2023

Efficient Bayesian Physics Informed Neural Networks for Inverse Problems via Ensemble Kalman Inversion
Andrew Pensoneault, Xueyu Zhu
High Dimensional Inverse Problem Hamiltonian Monte Carlo Physic Informed Machine Ensemble Kalman Inversion Bayesian Physic Informed Neural Network

November 1, 2022

A unified method of data assimilation and turbulence modeling for separated flows at high Reynolds numbers
Z. Y. Wang, W. W. Zhang
Unified Framework Data Assimilation General Flow Stratified Turbulence Turbulence Model Reynolds Number Ensemble Kalman Inversion

November 5, 2021

Reconstructing Training Data from Diverse ML Models by Ensemble Inversion
Qian Wang, Daniel Kurz
Generative Adversarial Network Training Data Machine Learning Model Model Inversion Ensemble Kalman Inversion

Ensemble Kalman Inversion

Papers

Reservoir History Matching of the Norne field with generative exotic priors and a coupled Mixture of Experts -- Physics Informed Neural Operator Forward Model

A Novel A.I Enhanced Reservoir Characterization with a Combined Mixture of Experts -- NVIDIA Modulus based Physics Informed Neural Operator Forward Model

Uncertainty quantification for deeponets with ensemble kalman inversion

Subsurface Characterization using Ensemble-based Approaches with Deep Generative Models

Gradient-free training of neural ODEs for system identification and control using ensemble Kalman inversion

Efficient Bayesian Physics Informed Neural Networks for Inverse Problems via Ensemble Kalman Inversion

A unified method of data assimilation and turbulence modeling for separated flows at high Reynolds numbers

Reconstructing Training Data from Diverse ML Models by Ensemble Inversion