I am interested in statistical machine learning and biologically-inspired computer vision, with an emphasis on unsupervised learning and time series analysis. Much of my work studies human movement.

I completed my PhD at the University of Toronto in 2009. My PhD thesis was titled "Composable, distributed-state models for high-dimensional time series". My thesis co-advisors were Geoffrey Hinton and Sam Roweis.

At NYU my faculty collaborators are Chris Bregler, Rob Fergus, and Yann LeCun.

Learn more about the crowd2cloud project. [Wired] [TV] [TV]

Adaptive Deconvolutional Networks for Mid and High Level Feature Learning
Matthew Zeiler, Graham Taylor, and Rob Fergus (2011)
To appear in Proc. of the 13th International Conference on Computer Vision (ICCV).

Learning Invarance through Imitation
Project page (with links to supplementary material)
Graham Taylor, Ian Spiro, Christoph Bregler, and Rob Fergus (2011)
Proc. of the 24th IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR).

Two Distributed-State Models for Generating High-Dimensional Time Series
Supplementary Material (including source code)
Graham Taylor, Geoffrey Hinton, and Sam Roweis (2011)
Journal of Machine Learning Research 12(Mar):1025--1068.
Note that this paper expands on two earlier works: (NIPS 2006), (ICML 2009).

Pose-Sensitive Embedding by Nonlinear NCA Regression
Supplementary Material (including source code)
Graham Taylor, Rob Fergus, George Williams, Ian Spiro and Christoph Bregler (2010)
Proc. of Advances in Neural Information Processing Systems (NIPS) 23.

Convolutional Learning of Spatio-temporal Features
Supplementary Material
Graham Taylor, Rob Fergus, Yann LeCun and Christoph Bregler (2010)
Proc. of the 11th European Conference on Computer Vision (ECCV).