[Ml-stat-talks] chong wang, wednesday december 7, 1pm
blei at CS.Princeton.EDU
Mon Dec 5 14:43:55 EST 2011
chong wang is giving his "pre-FPO" this wednesday at 1pm in CS 302.
(the preFPO is part of the computer science phd requirements. its a
public talk about 6m to 1y before graduation.)
please join us, especially those of you interested in probabilistic
models, bayesian nonparametrics, topic modeling, scalable computation,
and recommendation systems.
Chong Wang will present his preFPO on Wednesday December 7
at 1PM in Room 302 (note room!). The members of his committee are:
David Blei, advisor; Rob Schapire and Eric Xing (CMU), readers;
Christiane Fellbaum and Andrea LaPaugh, readers. Everyone is
invited to attend his talk. His abstract follows below.
Title: Hierarchical Bayesian modeling: models and inference algorithms
Hierarchical Bayesian models have become a popular tool for analyzing
large-scale real-world data, such as text and images. Through these
models, people can build useful tools for latent structure discovery,
browsing and recommendations. In this talk, I will present several
recent advances in the area of hierarchical Bayesian modeling, with an
emphasis on topic modeling, Bayesian nonparametrics and approximate
posterior inference. Specifically, I will talk about a novel model on
recommending scientific articles that provides an interpretable latent
structure for both users and items in a recommender system and then
present an online variational inference algorithm that scales up to
millions of articles on a single machine and also does model
selection. Finally, I will summarize with some ongoing work.
More information about the Ml-stat-talks