报告题目： Some New Bayesian Inference and Modeling Approaches to Time Series Data
This talk will discuss some new approaches to tracking and time series problems using ideas from Bayesian machine learning. The first part will focus on inference for the the classical nonlinear Kalman filtering problem. I will discuss new methods for posterior approximation based on the Kullback-Leibler divergence. The talk will then focus on a general, multiple time series modeling problem. I will discuss a new dynamic matrix factorization approach based on Bayesian nonparametrics, with an extension to deep modeling using the variational autoencoder.
John Paisley is an Associate Professor in the Department of Electrical Engineering at Columbia University, where he is also a member of the Data Science Institute. He received the B.S. and Ph.D. degrees in Electrical and Computer Engineering from Duke University in 2004 and 2010. Between 2010 and 2013 he was a postdoc in the Computer Science departments at Princeton University and UC Berkeley. His research focuses on Bayesian methods for machine learning, including Bayesian nonparametrics and variational inference techniques. He applies these techniques to several problems in signal and information processing, including compressed sensing and topic modeling.