A dimension-reduced variational approach for solving physics-based inverse problems using generative adversarial network priors and normalizing flows
Agnimitra Dasgupta, Dhruv Patel, Deep Ray, E. A. Johnson, Assad A. Oberai
Topics & Concepts
Posterior probabilityPrior probabilityMathematicsBayesian inferenceAlgorithmUncertainty quantificationMarkov chain Monte CarloFlow (mathematics)Inverse problemComputer scienceGenerative modelMathematical optimizationInferenceBayesian probabilityApplied mathematicsArtificial intelligenceMachine learningGenerative grammarMathematical analysisGeometryGenerative Adversarial Networks and Image SynthesisGaussian Processes and Bayesian InferenceModel Reduction and Neural Networks