This PIMS funded collaborative research
group focuses on Bayesian methods for network
analysis, paying special attention to model design and
computational issues of learning and inference.
Bayesian inference is an approach to statistics in
which all forms of uncertainty are expressed in terms
of probability. Non-Bayesian approaches to inference
have dominated statistical theory and practice for
most of the past century, but the last two decades
have seen a reemergence of Bayesian statistical
inference. This is mainly due to the dramatic increase
in computer power and the availability of new
computational tools, including variational techniques,
Markov chain Monte Carlo (MCMC) and sequential Monte
Carlo (SMC). Bayesian modeling has become common
practice as it provides a powerful method for coping
with very complex stochastic domains, including
networks. Networks are widely used to represent data
on relations between interacting actors or nodes.
Among many things, they can be used to describe social
networks, genetic regulatory networks, computer
networks, and sensor networks. In these settings,
traditional independence assumptions are blatantly
inappropriate; the structure of relationships between
the data must be taken into account. As a result,
there has been increasing research developing
techniques for incorporating network structures into
machine learning and statistics. This collaborative
research group will bring together researchers working
on Bayesian modeling for networks from different
communities, thereby fostering collaborations and
intellectual exchange. Our hope is that this will
result in novel modeling approaches, diverse
applications, and new research directions. In
particular we will focus on three main problems:
social networks, regulatory networks and sensor
networks. Even though the three problems share many
similar features, both in terms of modeling and
computation, they are usually treated separately.
CRG Leaders: Raphael Gottardo (UBC), Paul Gustafson (UBC), Lurdes Inoue (UW), Adrian Raftery (UW) and Tim Swartz (SFU)
Other faculty participating include: Derek Bingham (SFU), Bertrand
Clarke (UBC), Nando De Freitas (UBC), Adrian Dobra (UW), Arnaud Doucet (UBC), Paramjit Gill (UBC-O), Peter Hoff (UW), Kevin Murphy (UBC), Dale Schuurmans (UoA), Cory Burtz (UoR).
Postdoc and students founded by PIMS: Francois Caron (PIMS-CRNS PDF,
UBC), Kenneth Lo (UBC)
Visitors: Radu Craiu (Statistics, University
of Toronto) April 15-23, 2008
Radu gave a talk on "Learn from Thy Neighbour:
Parallel-Chain Adaptive MCMC" on April 15, click here for more details.
Sylvia Richardson (Imperial
College) October, 2010