Preconditioned training of normalizing flows for variational inference in inverse problems
arXiv:2101.03709
Abstract
Obtaining samples from the posterior distribution of inverse problems with expensive forward operators is challenging especially when the unknowns involve the strongly heterogeneous Earth. To meet these challenges, we propose a preconditioning scheme involving a conditional normalizing flow (NF) capable of sampling from a low-fidelity posterior distribution directly. This conditional NF is used to speed up the training of the high-fidelity objective involving minimization of the Kullback-Leibler divergence between the predicted and the desired high-fidelity posterior density for indirect measurements at hand. To minimize costs associated with the forward operator, we initialize the high-fidelity NF with the weights of the pretrained low-fidelity NF, which is trained beforehand on available model and data pairs. Our numerical experiments, including a 2D toy and a seismic compressed sensing example, demonstrate that thanks to the preconditioning considerable speed-ups are achievable compared to training NFs from scratch.
References in corpus (10)
- How transferable are features in deep neural networks?
- Density estimation using Real NVP
- An introduction to sampling via measure transport
- Task Agnostic Continual Learning Using Online Variational Bayes
- Invert to Learn to Invert
- Iterative Amortized Inference
- Faster Uncertainty Quantification for Inverse Problems with Conditional Normalizing Flows
- Deep Probabilistic Imaging: Uncertainty Quantification and Multi-modal Solution Characterization for Computational Imaging
- Learned imaging with constraints and uncertainty quantification
- Fully reversible neural networks for large-scale surface and sub-surface characterization via remote sensing