Gaussian Processes are a principled, practical, probabilistic approach to learning in flexible non-parametric models and have found numerous applications in regression, classification, unsupervised learning and reinforcement learning. Inference, learning and prediction can be done exactly on small data sets with Gaussian likelihood. In more realistic application with large scale data and more complicated likelihoods approximations are necessary. The variational framework for approximate inference in Gaussian processes has emerged recently as a highly effective and practical tool. I will review and demonstrate the capabilities of this framework applied to non-linear state space models.
Biography: Carl is the Professor of Machine Learning at Cambridge University, and currently head of the famed Laboratory for Computational and Biological Learning there. He is widely credited with popularising the notion of Gaussian processes in the machine learning community. His textbook on the topic, written with Chris Williams, occupies a prominent spot on many a desk in Tübingen and all over the world. He is also the Chairman of prowler.io, a wildly successful ML-startup in Cambridge.