www
/
www
8
0
Fork 0
Code Issues Pull Requests 1 Projects Releases Wiki Activity

adding Geoffrey Hinton talk

This commit is contained in:
Edgar Bering 2010-01-07 23:17:35 -05:00
parent ce1ede5352
commit 4f141ac2d9
1 changed files with 31 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
events.xml
View File

@ -3,6 +3,37 @@
<!DOCTYPE eventdefs SYSTEM "csc.dtd">
<eventdefs>
<!-- Winter 2010 -->
<eventitem date="2010-01-26" time="05:00 PM" room="DC1302" title="Deep learning with multiplicative interactions">
<short><p>Geoffrey Hinton, from the University of Toronto and the Canadian Institute for Advanced Research, will discuss some of his latest work in learning networks and artificial intelligence. The talk will be accessable, so don't hesitate to come out. More information about Dr. Hinton's research can be found on <a href="http://www.cs.toronto.edu/~hinton/">his website</a>.
</p></short>
<abstract><p>Deep networks can be learned efficiently from unlabeled data. The layers
of representation are learned one at a time using a simple learning
module, called a "Restricted Boltzmann Machine" that has only one layer
of latent variables. The values of the latent variables of one
module form the data for training the next module. Although deep
networks have been quite successful for tasks such as object
recognition, information retrieval, and modeling motion capture data,
the simple learning modules do not have multiplicative interactions which
are very useful for some types of data.
</p><p>The talk will show how a third-order energy function can be factorized to
yield a simple learning module that retains advantageous properties of a
Restricted Boltzmann Machine such as very simple exact inference and a
very simple learning rule based on pair-wise statistics. The new module
contains multiplicative interactions that are useful for a variety of
unsupervised learning tasks. Researchers at the University of Toronto
have been using this type of module to extract oriented energy from image
patches and dense flow fields from image sequences. The new module can
also be used to allow motions of a particular style to be achieved by
blending autoregressive models of motion capture data.
</p></abstract>
</eventitem>
<!-- Fall 2009 -->
<eventitem date="2009-12-05" time="6:30 PM" room="MC3036" edate="2009-12-05" etime="11:55 PM" title="The Club That Really Likes Dinner">
<short><p>Come on out to the club's termly end of term dinner, details in the abstract</p></short>