Statistics Seminar: Lancelot F. James (Hong Kong University of Science and Technology) -- Some aspects of Bayesian non-parametrics and Combinatorial Stochastic processes

ABSTRACT:

In this talk I will discuss, from my own perspective, some
technical aspects of the field of  Bayesian non-parametric statistics and
its interface with concepts arising in CSP. Part of my talk will be
tutorial in nature aimed at those at a graduate or higher level and will
also sprinkle in a few things I have been thinking about if time permits.


BIO---

Prof. Lancelot F. James is currently a Professor in the Business School at
the Hong Kong University of Science and Technology. He has been at HKUST
since 2001. Prior to that, he was an Assistant Professor in the School of
Engineering at the Johns Hopkins University in Baltimore. He is well known
for his work in Bayesian nonparametric statistics. Since the early 2000s,
he has advocated and developed ideas related to the usage of Chinese
restaurant processes, stick-breaking priors, and Pitman-Yor processes.
Since that time, these colourfully named processes (the latter two named
by James and his co-author Hemant Ishwaran in 2001) have played a major
role in the development of intricate applications in Bayesian Statistical
Machine Learning and Bayesian nonparametric methods in general, all of
which are pertinent to the analysis of Big Data. Prof. James is an elected
member of the International Statistical Institute. Since 2008 he is an
elected Fellow of the Institute of Mathematical Statistics where he is
cited for contributions to Bayesian nonparametric statistics, the
development of Poisson partition calculus for Levy processes, and for
dedicated service to IMS. Professor James was born on the island of
 Jamaica in the Caribbean Sea and grew up in Westchester County, New York.