BayesFlow: Learning complex stochastic models with invertible neural networks

ST Radev, UK Mertens, A Voss… - IEEE transactions on …, 2020 - ieeexplore.ieee.org
Estimating the parameters of mathematical models is a common problem in almost all
branches of science. However, this problem can prove notably difficult when processes and …

Coupling-based invertible neural networks are universal diffeomorphism approximators

T Teshima, I Ishikawa, K Tojo, K Oono… - Advances in …, 2020 - proceedings.neurips.cc
Invertible neural networks based on coupling flows (CF-INNs) have various machine
learning applications such as image synthesis and representation learning. However, their …

Derivative-informed neural operator: an efficient framework for high-dimensional parametric derivative learning

T O'Leary-Roseberry, P Chen, U Villa… - Journal of Computational …, 2024 - Elsevier
We propose derivative-informed neural operators (DINOs), a general family of neural
networks to approximate operators as infinite-dimensional map**s from input function …

Coupling techniques for nonlinear ensemble filtering

A Spantini, R Baptista, Y Marzouk - SIAM Review, 2022 - SIAM
We consider filtering in high-dimensional non-Gaussian state-space models with intractable
transition kernels, nonlinear and possibly chaotic dynamics, and sparse observations in …

Sampling in unit time with kernel fisher-rao flow

A Maurais, Y Marzouk - arxiv preprint arxiv:2401.03892, 2024 - arxiv.org
We introduce a new mean-field ODE and corresponding interacting particle systems (IPS)
for sampling from an unnormalized target density. The IPS are gradient-free, available in …

Certified dimension reduction in nonlinear Bayesian inverse problems

O Zahm, T Cui, K Law, A Spantini, Y Marzouk - Mathematics of Computation, 2022 - ams.org
We propose a dimension reduction technique for Bayesian inverse problems with nonlinear
forward operators, non-Gaussian priors, and non-Gaussian observation noise. The …

On the representation and learning of monotone triangular transport maps

R Baptista, Y Marzouk, O Zahm - Foundations of Computational …, 2024 - Springer
Transportation of measure provides a versatile approach for modeling complex probability
distributions, with applications in density estimation, Bayesian inference, generative …

Conditional sampling with monotone GANs: From generative models to likelihood-free inference

R Baptista, B Hosseini, NB Kovachki… - SIAM/ASA Journal on …, 2024 - SIAM
We present a novel framework for conditional sampling of probability measures, using block
triangular transport maps. We develop the theoretical foundations of block triangular …

A dimension-reduced variational approach for solving physics-based inverse problems using generative adversarial network priors and normalizing flows

A Dasgupta, DV Patel, D Ray, EA Johnson… - Computer Methods in …, 2024 - Elsevier
We propose a novel modular inference approach combining two different generative models—
generative adversarial networks (GAN) and normalizing flows—to approximate the posterior …

[HTML][HTML] Efficiency comparison of MCMC and Transport Map Bayesian posterior estimation for structural health monitoring

J Grashorn, M Broggi, L Chamoin, M Beer - Mechanical Systems and Signal …, 2024 - Elsevier
In this paper, an alternative to solving Bayesian inverse problems for structural health
monitoring based on a variational formulation with so-called transport maps is examined …