This is to advise you that the upcoming MITACS seminar has been rescheduled for Friday, September 18th from 11:00-12:00, due to some scheduling conflicts. I apologize for the short notice and hope many of you can make it.

---------- Forwarded message ---------- From: Guillaume Desjardins guillaume.desjardins@gmail.com Date: Tue, Sep 8, 2009 at 5:36 PM Subject: UdeM-McGill-MITACS machine learning seminar (NEW DATE & TIME) Mon, Sept 18th @11h00 To: lisa_seminaires@iro.umontreal.ca

We start off the new academic year with a talk from Dr. Massih-Reza Amini entitled "Learning from Multiple Partially Observed Views".

Speaker: Dr. Massih-Reza Amini              Associate Professor of  Pierre & Marie Curie University /              Research Officer, NRC Institute for Information Technology

Location: Pavillon André-Aisenstadt (UdeM), room 3195 Time: September 18th 2009 at 11h00

Title: Learning from Multiple Partially Observed Views

Résumé: We address the problem of learning classifiers when observations have multiple views, some of which may not be observed for all examples. We assume the existence of view generating functions which may complete the missing views in an approximate way. This situation corresponds for example to learning text classifiers from multilingual collections where documents are not available in all languages. In that case, Machine Translation (MT) systems may be used to translate each document in the missing languages. We derive a generalization error bound for classifiers learned on examples with multiple artificially created views. Our result uncovers a trade-off between the size of the training set, the number of views, and the quality of the view generating functions. As a consequence, we identify situations where it is more interesting to use multiple views for learning instead of classical single view learning. An extension of this framework is a natural way to leverage unlabeled multi-view data in semi-supervised learning. Experimental results on a subset of the Reuters RCV1/RCV2 collections support our findings by showing that additional views obtained from MT may significantly improve the classification performance in the cases identified by our trade-off.