Neural Network Research Group
Our group invents and evaluates novel learning algorithms for neural networks.
Our main aims are to gain insight into how the brain learns, to develop
algorithms that are useful for practical tasks, and to understand the
relationship of neural network algorithms to other statistical approaches.
Projects
- Helmholtz machines: Using bottom-up connections to fit top-down models
- Mixture models: Modelling complex distributions as mixtures of simple ones
- Ensemble learning: Fitting weight distributions without Monte Carlo
- Bayesian inference: Making predictions using all likely networks, not just one
- Monte Carlo methods: Solving hard Bayesian inference problems stochastically
- Elastic models: Using deformable models to recognize hand-written digits
- GloveTalk: Using neural networks to convert gestures to real-time speech
- Xerion: Unix software for neural network simulation
- Delve: Data and software for evaluating learning algorithms
People
- Faculty Members:
Geoffrey Hinton,
Radford Neal,
Rob Tibshirani
- Visiting Researchers:
David MacKay,
Jim Stone
- Postdoctoral Fellows:
Zoubin Ghahramani
- Graduate Students:
Andy Brown, Xiao-Peng Cheng,
Brendan Frey,
Sageev Oore
Alberto Paccanaro, Brian Sallans
- Staff:
Michael Revow (Research associate),
Maureen Smith (Administrative assistant)
- Former Group Members:
Sue Becker,
Michelle Craig, Yann le Cun, Peter Dayan,
Sidney Fels, Conrad Galland, Mike Mozer, Steve Nowlan,
Tony Plate,
Carl Edward Rasmussen,
Virginia de Sa,
Evan Steeg,
Ava To,
Ray Watrous, Lisa White,
Chris Williams,
Rich Zemel.
Address
Department. of Computer Science
University of Toronto
6 King's College Road
Toronto, Ontario M5S 3H5
CANADA
Maureen's phone: +1 416-978-7403
Neuron lab phone: +1 416-978-7391
FAX: +1 416-978-1455