Dr Simon Osindero, PhD
E m a i l : o s i n d e r o @ c s . t o r o n t o . e d u
My work currently spans a number of areas. These are mostly within the fields of machine learning, bayesian statistics, computer vision, computational biology and computational linguistics.
I am particularly inspired by work at the interface of computer science and neuroscience. I believe that concepts from machine learning and information theory are vital if we wish to develop a solid discipline of theoretical neurobiology. Concurrently, insights from biology can help us design better artificial intelligence systems.
At present one of my main projects involves developing models and algorithms to perform learning and inference in probabilistic hierarchical representations. Such models hold great promise for practical applications such as machine vision and image processing, as well as providing a basis for understanding neural coding.
- Modelling Image Patches With A Directed Hierarchy Of Markov Random Fields
Osindero, S. and Hinton, G. E.
Advances In Neural Information Processing Systems,20, 2008 pdf
- Combining Discriminative Features To Infer Complex Trajectories
Ross, D., Osindero, S. and Zemel, R.
Proceedings of the 23rd International Conference on Machine Learning, 2006 ps pdf
- A Fast Learning Algorithm For Deep Belief Networks
Hinton, G. E., Osindero, S. and Teh, Y. W.
Neural Computation, 18 (7), 2006 pdf
- Unsupervised Discovery of Non-linear Structure Using Contrastive Backpropagation
Hinton, G. E., Osindero, S., Welling, M. and Teh, Y. W.
Cognitive Science, 30(4), 2006 pdf
- An Alternative Infinite Mixture of Gaussian Process Experts
Meeds, E. and Osindero, S.
Advances In Neural Information Processing Systems,18, 2006 ps pdf
- Topographic Product Models Applied To Natural Scene Statistics
Osindero, S., Welling, M. and Hinton, G. E.
Neural Computation, 18 (2), 2006 pdf
Learning Causally Linked Markov Random Fields
Hinton, G. E., Osindero, S. and Bao, K.
Proceedings of the 10th International Workshop on Artificial Intelligence and Statistics, 2005 ps pdf
Energy-Based Models for Sparse Overcomplete Representations
Teh, Y. W, Welling, M., Osindero, S. and Hinton G. E.
Journal of Machine Learning Research, 4, 2003 pp 1235-1260. ps pdf
Learning Sparse Topographic Representations with Products of Student-t Distributions
Welling, M., Hinton, G. E. and Osindero, S.
Advances in Neural Information Processing Systems, 15, 2003, MIT Press, Cambridge, MA ps
A New View of ICA
Hinton G. E., Welling, M., Teh, Y. W, and Osindero, S
Proceedings of ICA-2001, San Diego, CA. ps pdf