Approximate Inference; Convolutional Neural Networks; Stein's Method
Approximate inference algorithms [Slides] Souvik Chakraborty Convolutional Neural Networks [Slides] Navid Shervani-Tabar Intro to Stein's method, with applications to Bayesian Surrogate Modeling [Slides] Yinhao Zhu Stein's method [1] is a theoretical technique to obtain bounds on the distances between probability distributions. It was recently combined with kernel methods and applied to Bayesian inference [2]. This talk reviews the basics of this recent development and shows
Parallel approaches for the Bayesian Gaussian process latent variable model; Regularization in optim
Parallel approaches for the Bayesian Gaussian process latent variable model [Slides] Steven Atkinson We consider the task of training a Bayesian GP-LVM and uncover opportunities to employ parallelism in computing the collapsed lower bound to the model evidence and its gradients with respect to its variational parameters. We will discuss its implementation within a C++ code for deep GPs. Reference: Gal, Yarin, Mark van der Wilk, and Carl Edward Rasmussen. "Distributed variati
