Condensed Stein Variational Gradient Descent for Uncertainty Quantification of Neural Networks

Useful for Phase estimation beyond the shot noise limit using Bayesian statistics. Based on Stein’s Lemma.