Efficient Posterior Inference & Generalization in Physics-Based Bayesian Inference with Conditional GANs

Abstract

In this work, we propose a conditional generative adversarial network (cGAN) to sample from the posterior of physics-based Bayesian inference problems. We utilize a U-Net architecture for the generator and inject the latent variable using conditional instance normalization. We solve the inverse heat conduction problem and demonstrate how the proposed strategy effectively quantifies the uncertainty in the inferred field. We also show that the structure of the generator promotes generalizability due to the local-nature of the learned inverse map.

PDF NeurIPSW OpenReview Semantic Scholar

Cite

Text

Ray et al. "Efficient Posterior Inference & Generalization in Physics-Based Bayesian Inference with Conditional GANs." NeurIPS 2021 Workshops: Deep_Inverse, 2021.

Markdown

[Ray et al. "Efficient Posterior Inference & Generalization in Physics-Based Bayesian Inference with Conditional GANs." NeurIPS 2021 Workshops: Deep_Inverse, 2021.](https://mlanthology.org/neuripsw/2021/ray2021neuripsw-efficient/)

BibTeX

@inproceedings{ray2021neuripsw-efficient,
  title     = {{Efficient Posterior Inference & Generalization in Physics-Based Bayesian Inference with Conditional GANs}},
  author    = {Ray, Deep and Patel, Dhruv V and Ramaswamy, Harisankar and Oberai, Assad},
  booktitle = {NeurIPS 2021 Workshops: Deep_Inverse},
  year      = {2021},
  url       = {https://mlanthology.org/neuripsw/2021/ray2021neuripsw-efficient/}
}