OCTOBER 6, 2009
DR. JONATHAN SCHAEFFER, University of Alberta
“Computer (and Human) Perfection at Checkers”
In 1989 the Chinook project began with the goal of winning the human World Checkers Championship. There was an imposing obstacle to success -- the human
champion, Marion Tinsley. Tinsley was as close to perfection at the game as was humanly possible. To be better than Tinsley meant that the computer had to be
perfect. In effect, one had to solve checkers. Little did we know that our quest would take 18 years to complete. In this talk, the creator of Chinook tells the story of the quest for computer and human perfection at the game of checkers.
OCTOBER 20, 2009
DR. GEOFFREY HINTON, University of Toronto
“Deep Learning for Object Recognition and Motion Generation”
I will describe how multiple layers of non-linear features can be learned one layer at a time from unlabeled images. The multiple layers of features form a good
generative model of images and the highest level features are very good for object recognition. I will show how the same learning methods can be applied to
high-dimensional sequential data to produce good generative models of people walking in many different styles.
NOVEMBER 3, 2009
DR. SIMON HAYKIN, McMaster University
“Cognitive Dynamic Systems”
This lecture will describe a new generation of engineering systems with cognition as the enabler. I will begin by describing the perception-action cycle that is basic to
the visual brain. Then I will demonstrate how cognitive information (signal) processing is so basic to the underlying theory and design of: cognitive radio, cognitive
radar, and cognitive mobile assistants which constitute the three pillars of my research program. I will finish the lecture by doing two things: present new and exciting
results on cognitive tracking radar and thereby demonstrate the power of cognition; and describe my vision on future research on Cognitive Dynamic Systems.
NOVEMBER 17, 2009
DR. RICHARD VAUGHAN, Simon Fraser University
“World-Mediated Robot Intelligence”
Unlike disembodied AI programs, robots are embedded in the same physical world as humans and other animals. Like animals, they must act appropriately - we say
"with intelligence" - to achieve their goals. The real world presents problems of uncertainty and the danger of running out of energy. Yet the world presents resources
that can help the robot, such as other agents. This talk discusses robot systems that sense and exploit regularities in the behaviour of other robots and animals to
obtain energy and work.
DECEMBER 1, 2009
DR. THOMAS TRAPPENBERG, Dalhousie University
“Real Intelligence: The Anticipating Brain”
The area of AI was always inspired by the human mind. I will try to give some perspective of new directions in AI from new hypothesis about cognitive processes and
neuroscience. Huge progress has been made in the scientific areas of neuroscience on one side and machine learning on the other, but it seems that these areas
developed largely independently since the exciting days of the perceptron half a century ago. However, there is now some exciting new convergence of these areas of
research. In particular, generative systems have made a strong impact on machine learning, and probabilistic reasoning replaced most of traditional expert-system
approaches. This seminar will explore related new developments in computational neuroscience, specifically systems with a large top-down component that are
capable of learning to anticipate the world. We further discuss representations of uncertainties in the brain and how synaptic mechanisms may contribute in such
IRMACS and AARMS host the Coast to Coast Seminar Series in partnership with ACEnet, SHARCNET, SCInet and WestGrid. This is a series of talks covering topics in mathematics and computer science, presented by researchers across Canada. These seminars are typically an hour in length and are presented using Access Grid videoconferencing facilities. To attend these seminars in person, please refer to the participating locations listed on our Seminar Locations page. If you have any questions or would like more information about this seminar series, please email email@example.com.
If you have any speaker or seminar topic suggestions, please send them to firstname.lastname@example.org.