Stochastic normalizing flows for inverse problems: a Markov chains viewpoint
From MaRDI portal
Applications of Markov chains and discrete-time Markov processes on general state spaces (social mobility, learning theory, industrial processes, etc.) (60J20) Computational methods in Markov chains (60J22) Bayesian inference (62F15) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Artificial neural networks and deep learning (68T07)
Abstract: To overcome topological constraints and improve the expressiveness of normalizing flow architectures, Wu, K"ohler and No'e introduced stochastic normalizing flows which combine deterministic, learnable flow transformations with stochastic sampling methods. In this paper, we consider stochastic normalizing flows from a Markov chain point of view. In particular, we replace transition densities by general Markov kernels and establish proofs via Radon-Nikodym derivatives which allows to incorporate distributions without densities in a sound way. Further, we generalize the results for sampling from posterior distributions as required in inverse problems. The performance of the proposed conditional stochastic normalizing flow is demonstrated by numerical examples.
Recommendations
- Stochastic normalizing flows as non-equilibrium transformations
- Normalizing Flows Aided Variational Inference: A Useful Alternative to MCMC?
- Solving inverse problems in stochastic models using deep neural networks and adversarial training
- Solving Bayesian inverse problems from the perspective of deep generative networks
- Stein variational gradient descent on infinite-dimensional space and applications to statistical inverse problems
Cites work
- BAYESIAN APPROACH TO THE STATISTICAL INVERSE PROBLEM OF SCATTEROMETRY: COMPARISON OF THREE SURROGATE MODELS
- Bayesian Gaussian mixture linear inversion for geophysical inverse problems
- Brownian motion, martingales, and stochastic calculus
- Exponential convergence of Langevin distributions and their discrete approximations
- General state space Markov chains and MCMC algorithms
- scientific article; zbMATH DE number 7370595 (Why is no real title available?)
- Invertible neural networks versus MCMC for posterior reconstruction in grazing incidence X-ray fluorescence
- Iterative multiplicative filters for data labeling
- On the convergence of the Metropolis-Hastings Markov chains
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Solving inverse problems using data-driven models
- Stabilizing Invertible Neural Networks Using Mixture Models
Cited in
(24)- Stochastic normalizing flows as non-equilibrium transformations
- Normalizing field flows: solving forward and inverse stochastic differential equations using physics-informed flow models
- Normalizing flows for probabilistic modeling and inference
- Generalized Normalizing Flows via Markov Chains
- Normalizing Flows Aided Variational Inference: A Useful Alternative to MCMC?
- Training normalizing flows with computationally intensive target probability distributions
- Variational inference in neural functional prior using normalizing flows: application to differential equation and operator learning problems
- Self-Supervised Deep Learning for Image Reconstruction: A Langevin Monte Carlo Approach
- Reduced-order model-based variational inference with normalizing flows for Bayesian elliptic inverse problems
- WPPNets and WPPFlows: The Power of Wasserstein Patch Priors for Superresolution
- NF-ULA: normalizing flow-based unadjusted Langevin algorithm for imaging inverse problems
- Less interaction with forward models in Langevin dynamics: enrichment and homotopy
- Learning variational autoencoders via MCMC speed measures
- Multi-resolution continuous normalizing flows
- Continuous Generative Neural Networks: A Wavelet-Based Architecture in Function Spaces
- Learning from small data sets: patch-based regularizers in inverse problems for image reconstruction
- Neural-network-based regularization methods for inverse problems in imaging
- Conditional Wasserstein distances with applications in Bayesian OT flow matching
- Sampling from Boltzmann densities with physics informed low-rank formats
- Multilevel diffusion: infinite dimensional score-based diffusion models for image generation
- Stability of data-dependent ridge-regularization for inverse problems
- A constraint-preserving neural network approach for mean-field games equilibria
- A mathematical explanation of transformers
- Neural Sampling with Reweighted Normalizing Flows via the Wasserstein--Fisher--Rao JKO Scheme
This page was built for publication: Stochastic normalizing flows for inverse problems: a Markov chains viewpoint
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5052899)
