Campus Event Calendar
Event Entry
What and Who
Approximate Submodular k-Partitions: Application to SNP Prediction
Dr. Jeff Bilmes
University of Washington, Seattle and MPI für Kybernetik, Tübingen
Talk
AG 1, AG 3, AG 4, AG 5, SWS, RG1, MMCI
Expert Audience
English
Note: We use this to send email in the morning.
Date, Time and Location
Monday, 18 May 2009
14:00
60 Minutes
E1 4
019
Saarbrücken
Abstract
Submodular functions are an important concept in combinatorial
optimization that have recently received interest by the machine
learning community. In this talk we will show how it is possible in
polynomial time to construct an approximately optimal $k$-partition of
a submodular function using a Gomory-Hu tree. We will see how this
yields a clustering algorithm on the multi-information function (the
generalization of mutual-information). We apply this algorithm to the
task of predicting single nucleotide polymorphisms (SNPs) and show
results that suggest state-of-the-art performance on this task.
Given time, we will also briefly mention other activities going on at
University of Washington, Seattle and MPI.
optimization that have recently received interest by the machine
learning community. In this talk we will show how it is possible in
polynomial time to construct an approximately optimal $k$-partition of
a submodular function using a Gomory-Hu tree. We will see how this
yields a clustering algorithm on the multi-information function (the
generalization of mutual-information). We apply this algorithm to the
task of predicting single nucleotide polymorphisms (SNPs) and show
results that suggest state-of-the-art performance on this task.
Given time, we will also briefly mention other activities going on at
University of Washington, Seattle and MPI.
Contact
Matthias Seeger
452
--email hidden
passcode not visible
logged in users only
Sabine Budde, 05/13/2009 16:40 -- Created document.