I am a PhD student in the Department of Computer Science at the University of Toronto. I am co-supervised by Geoffrey Hinton and Sam Roweis. I am interested in Machine Learning, as well as other aspects of the broader world of Artificial Intelligence. In the past, I have worked extensively in Reinforcement Learning, applied to Computer Vision.

This winter I will work on research and TA CSC321 (Machine Learning and Neural Networks).

ICASSP 2008

Jan 21, 2008 16:52

A paper related to the work I did while at MSR has been accepted to ICASSP 2008 (Las Vegas!).
Update: Download the paper

Fall term

Sep 4, 2007 15:13

I just finished a summer internship at Microsoft Research (Redmond) where I worked with Mike Seltzer.

NESCAI this weekend

Apr 12, 2007 22:39

I'll be at the NESCAI conference this weekend in Ithaca, NY. I've added some demos here showing the same model we've used for modeling human motion generating "video texture".

NIPS paper updated

Jan 13, 2007 9:42

The final version of our paper "Modeling Human Motion Using Binary Latent Variables" is now available here.

NIPS

Sep 26, 2006 10:27

A paper I have recently co-authored with Geoff Hinton and Sam Roweis has been accepted for a poster spotlight at NIPS 19 in December. This paper proposes a new model for human motion (mocap) data which uses a large number of hidden binary variables that are symmetrically connected to the observable variables to form a restricted Boltzmann machine. In addition, the hidden variables are linked together by directed connections to form a sigmoid belief net. This paper describes how the model is trained, and demonstrates that it can generate realistic motion.

More information on this project, including a preprint is available here.

Graham Taylor
Graduate Student
Department of Computer Science
University of Toronto
10 King's College Rd.
Room 3302
Toronto, Ontario, Canada
M5S 3G4

If you would like to email me, please email gwtaylor at the domain you see above (everything minus the www).

Nothing here at the moment.