Title
Normalized Random Measures and Extensions
Abstract
The normalized random measure is a large class of random probability measures that has appealing properties and wide applications in Bayesian nonparametrics. In this talk I will review the Poisson process, completely random measures, and how to construct normalized random measure from them. I will then talk about its posterior characterization,
and how to construct dependent normalized random measures from it, e.g., spatial normalized random measures and mixed normalized random measures. I then will show the applications of the dependent normalized random measures in topic models. Some other potential applications will be discussed as well.
About the presenter
Changyou is currently a PhD candidate in the College of Engineering and Computer Science, the Australian National University. He is also affiliated with the Statistical Machine Learning (SML) group at National ICT Australia (NICTA). His primary supervisor is Prof. Wray Buntine.
He received his bachelor and master degrees in 2007 and 2010,respectively, both from the School of Computer Science, Fudan University, Shanghai, China, supervised by Prof. Junping Zhang.