KRNet Map

KRNet, and its bounded variant B-KRNet, are invertible mapping techniques used primarily for density estimation and approximation, particularly within high-dimensional spaces. Research focuses on adapting KRNet architectures, such as incorporating Kalman filters for improved performance in autonomous driving or employing variational autoencoders for dimension reduction in Bayesian inverse problems. These methods offer efficient solutions for handling complex probability distributions, finding applications in diverse fields ranging from autonomous systems to solving partial differential equations.