Print Friendly printer friendly
 Events
Yariv Mizrahi
Thu 6 Nov 2014, 12:30pm SPECIAL
One Time Event
Room 200 of the Graduate Student Centre
Doctoral Exams
Room 200 of the Graduate Student Centre
Thu 6 Nov 2014, 12:30pm-2:30pm

Details

ABSTRACT We introduce a new class of parallel parameter learning algorithms for Markov random fields (MRFs) with untied parameters, which are efficient for a large class of practical models. The algorithms parallelize naturally over cliques and, for graphs of bounded degree, have complexity that is linear in the number of cliques. We refer to these algorithms with the acronym LAP, which stands for Linear And Parallel. Unlike their competitors, the marginal versions of the proposed algorithms are fully parallel and for log-linear models they are also data efficient, requiring only the local sufficient statistics of the data to estimate parameters. LAP algorithms are ideal for parameter learning in big graphs and big data applications. The correctness of the newly proposed algorithms relies heavily on the existence and uniqueness of the normalized potential representation of an MRF. We capitalize on this theoretical result to develop a new theory of correctness and consistency of LAP estimators corresponding to different local graph neighborhoods. This theory also establishes a general condition on composite likelihood decompositions of MRFs that guarantees the global consistency of distributed estimators, provided the local estimators are consistent. We introduce a conditional variant of LAP that enables us to attack parameter estimation of fully observed models of arbitrary connectivity, including fully connected Boltzmann distributions. We show consistency for this distributed estimator, and relate it to distributed pseudo-likelihood estimators. Finally, for linear and non-linear inverse problems with a sparse forward operator, we present a new algorithm, named iLAP, which decomposes the inverse problem into a set of smaller dimensional inverse problems that can be solved independently.
hide
Deanna Needell
Thu 13 Nov 2014, 12:00pm
One Time Event
4133 ESB (PIMS lounge)
TBA
4133 ESB (PIMS lounge)
Thu 13 Nov 2014, 12:00pm-1:00pm

Details

 
hide
Tyler Helmuth
Tue 25 Nov 2014, 4:00pm SPECIAL
One Time Event
Graduate Student Center, Room 203
Doctoral Exam
Graduate Student Center, Room 203
Tue 25 Nov 2014, 4:00pm-6:30pm
hide
Sat 10 Jan 2015, 9:00am SPECIAL
One Time Event
To Be Announced
Analysis - Qualifying Exams
To Be Announced
Sat 10 Jan 2015, 9:00am-12:00pm
hide
Sat 10 Jan 2015, 1:00pm SPECIAL
One Time Event
To Be Announced
Differential Equations - Qualifying Exams
To Be Announced
Sat 10 Jan 2015, 1:00pm-4:00pm
hide
Sat 10 Jan 2015, 1:00pm SPECIAL
One Time Event
To Be Announced Later
Algebra Qualifying Exams
To Be Announced Later
Sat 10 Jan 2015, 1:00pm-4:00pm
hide
 
Top