ASPIRE: Iterative Amortized Posterior Inference for Bayesian Inverse Problems

arXiv:2405.05398v1 Announce Type: new
Abstract: Due to their uncertainty quantification, Bayesian solutions to inverse problems are the framework of choice in applications that are risk averse. These benefits come at the cost of computations that are in general, intractable. New advances in machine learning and variational inference (VI) have lowered the computational barrier by learning from examples. Two VI paradigms have emerged that represent different tradeoffs: amortized and non-amortized. Amortized VI can produce fast results but due to generalizing to …

abstract advances applications arxiv aspire bayesian benefits cost cs.lg framework general inference iterative machine machine learning posterior quantification risk solutions stat.ml type uncertainty