Abstract:We propose a novel modular inference approach combining two different generative models — generative adversarial networks (GAN) and normalizing flows — to approximate the posterior distribution of physics-based Bayesian inverse problems framed in high-dimensional ambient spaces. We dub the proposed framework GAN-Flow. The proposed method leverages the intrinsic dimension reduction and superior sample generation capabilities of GANs to define a low-dimensional data-driven prior distribution. Once a trained GAN-prior is available, the inverse problem is solved entirely in the latent space of the GAN using variational Bayesian inference with normalizing flow-based variational distribution, which approximates low-dimensional posterior distribution by transforming realizations from the low-dimensional latent prior (Gaussian) to corresponding realizations of a low-dimensional variational posterior distribution. The trained GAN generator then maps realizations from this approximate posterior distribution in the latent space back to the high-dimensional ambient space. We also propose a two-stage training strategy for GAN-Flow wherein we train the two generative models sequentially. Thereafter, GAN-Flow can estimate the statistics of posterior-predictive quantities of interest at virtually no additional computational cost. The synergy between the two types of generative models allows us to overcome many challenges associated with the application of Bayesian inference to large-scale inverse problems, chief among which are describing an informative prior and sampling from the high-dimensional posterior. GAN-Flow does not involve Markov chain Monte Carlo simulation, making it particularly suitable for solving large-scale inverse problems. We demonstrate the efficacy and flexibility of GAN-Flow on various physics-based inverse problems of varying ambient dimensionality and prior knowledge using different types of GANs and normalizing flows. Notably, one of the applications we consider involves a 65,536-dimensional inverse problem of phase retrieval wherein an object is reconstructed from sparse noisy measurements of the magnitude of its Fourier transform.

An Adversarial Non-Volume Preserving Flow Model with Boltzmann Priors

VFlow: More Expressive Generative Flows with Variational Data Augmentation

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

AmbientFlow: Invertible generative models from incomplete, noisy measurements

Variational Potential Flow: A Novel Probabilistic Framework for Energy-Based Generative Modelling

Convergence of flow-based generative models via proximal gradient descent in Wasserstein space

RG-Flow: A hierarchical and explainable flow model based on renormalization group and sparse prior

Monge-Amp\`ere Flow for Generative Modeling

A Tale of Two Latent Flows: Learning Latent Space Normalizing Flow with Short-run Langevin Flow for Approximate Inference

C-Flow: Conditional Generative Flow Models for Images and 3D Point Clouds

Text-to-Image Rectified Flow as Plug-and-Play Priors

Flow Priors for Linear Inverse Problems via Iterative Corrupted Trajectory Matching

Monge-Ampère Flow for Generative Modeling

Variational Probability Flow for Biologically Plausible Training of Deep Neural Networks.

Flow Matching in Latent Space

Deep Generative Learning via Variational Gradient Flow

Verlet Flows: Exact-Likelihood Integrators for Flow-Based Generative Models

Variational Flow Models: Flowing in Your Style

Improved variational inference with inverse autoregressive flow

PSRFlow: Probabilistic Super Resolution with Flow-Based Models for Scientific Data