THE 2010 WINTER SCHOOL ON
MACHINE LEARNING AND COMPUTER VISION

Indian Institute of Science Campus, Bangalore, India
January 9-17, 2010

Sponsored by
Canadian Institute for Advanced Research (CIFAR)
Microsoft Research India (MSR India)

Speakers

Description

Machine learning and its application to visual data processing is one of the most interesting, interdisciplinary and rapidly growing areas of research. The objective of this winter school is to equip graduate students, faculty members and outstanding undergraduate students with a variety of tools needed to conduct research in this area.

The purpose of a learning algorithm is to take training data as input and construct a model that: (a) accounts for the training data by identifying combinations of patterns and (b) is useful for making predictions for previously unseen data. A simple example is clustering where, for example, we might want to group together all the visually similar images returned by a search and then display only the representative image from each cluster. Important research considerations include how to automatically determine the number of clusters, how to learn an appropriate measure of similarity, how to determine exemplars and cluster membership, etc. A more difficult example is learning models of objects from a training set of videos, and then explaining a new video as a composition of those objects, their parts, their motion and their interactions.

Learning algorithms can be understood using concepts from many areas including functional analysis, convex analysis and optimization, message-passing, variational techniques and Monte Carlo methods. The participants will be taught some of these techniques, but should have background in undergraduate calculus, linear algebra, probability and statistics. Background in optimization and signal processing would be helpful but is not required.

National and international leaders in the field (see the above list) will give introductory tutorials, advanced seminars and will assist in small-group discussions. Laboratory sessions will give participants an opportunity to implement and explore the taught material for themselves. The first two days of the school will consist of introductory tutorials, while the following seven days will include more advanced tutorials plus research seminars.

Schedule

(to be updated)

How to attend the Winter School

The last for applying for the Winter School has passed. The organizing committee shall not accept any more applications. All applicants will be notified of their selection status as per the date mentioned below.

Important Dates

Announcement of selected attendees: Monday, November 16th 2009
Winter School dates: Saturday, January 9th 2010 to Sunday, January 17th 2010

Please Note:

• CIFAR and Microsoft Research India shall provide for travel, lodging and boarding of all attendees that come from within India, outside of Bangalore, for the Winter School.
• There are no fees that attendees have to pay for the Winter School.
• Attendees from outside India are welcome to apply though CIFAR and Microsoft Research India will only provide them accommodation in Bangalore and no travel funds.
• For any enquiries, please email msriwsap@microsoft.com.

Organization

Co-Chairs

Local Organizing Committee

Support from CIFAR and MSR India

Acknowledgements

The organizers would like to acknowledge additional support provided by the Indian Institute of Science, Bangalore, and the Department of Electrical and Computer Engineering at the University of Toronto.