Geoffrey E. Hinton

I now work part-time for Google as an Engineering Fellow and part-time for the University of Toronto as an Emeritus Distinguished Professor. For much of the year, I work at the University in the morning and at the Google Toronto office at 111 Richmond Street from 2.00pm to 6.00pm. I also spend several months per year working full-time for Google in Mountain View, California.

Department of Computer Science   email: hinton [at] cs [dot] toronto [dot] edu
University of Toronto   voice: 416-978-7564
6 King's College Rd.   fax: 416-978-1455
Toronto, Ontario   office: Pratt 290G
M5S 3G4, CANADA   Directions for Visitors
 

Check out the new web page for Machine Learning at Toronto

Information for prospective students:
I will not be taking any more visiting students, summer students or visitors. I will not be the sole advisor of any new graduate students, but I may co-advise some new graduate students with other U of T machine learning faculty. I also advise some of the residents in the Google Brain Residents Program

News
Results of the 2012 competition to recognize 1000 different types of object
How George Dahl won the competition to predict the activity of potential drugs
How Vlad Mnih won the competition to predict job salaries from job advertisements
href="http://blog.kaggle.com/2012/11/02/t-distributed-stochastic-neighbor-embedding-wins-merck-viz-challenge/"> How Laurens van der Maaten won the competition to visualize a dataset of potential drugs

Basic papers on deep learning

Hinton, G. E., Osindero, S. and Teh, Y. (2006)
A fast learning algorithm for deep belief nets.
Neural Computation, 18, pp 1527-1554. [pdf]
Movies of the neural network generating and recognizing digits

Hinton, G. E. and Salakhutdinov, R. R. (2006)
Reducing the dimensionality of data with neural networks.
Science, Vol. 313. no. 5786, pp. 504 - 507, 28 July 2006.
[ full paper ] [ supporting online material (pdf) ] [ Matlab code ]

LeCun, Y., Bengio, Y. and Hinton, G. E. (2015)
Deep Learning
Nature, Vol. 521, pp 436-444. [pdf]

Papers on deep learning without much math

Hinton, G. E. (2007)
To recognize shapes, first learn to generate images
In P. Cisek, T. Drew and J. Kalaska (Eds.)
Computational Neuroscience: Theoretical Insights into Brain Function. Elsevier. [pdf of final draft]

Hinton, G. E. (2007)
Learning Multiple Layers of Representation.
Trends in Cognitive Sciences, Vol. 11, pp 428-434. [pdf]

Hinton, G. E. (2014)
Where do features come from?.
Cognitive Science, Vol. 38(6), pp 1078-1101. [pdf]

A practical guide to training restricted Boltzmann machines
[pdf]

Recent Papers

LeCun, Y., Bengio, Y. and Hinton, G. E. (2015)
Deep Learning
Nature, Vol. 521, pp 436-444. [pdf]

Hinton, G. E., Vinyals, O., and Dean, J. (2015)
Distilling the knowledge in a neural network.
arXiv preprint arXiv:1503.02531 [pdf]

Vinyals, O., Kaiser, L., Koo, T., Petrov, S., Sutskever, I., & Hinton, G. E. (2014)
Grammar as a foreign language.
arXiv preprint arXiv:1412.7449 [pdf]

Hinton, G. E. (2014)
Where do features come from?.
Cognitive Science, Vol. 38(6), pp 1078-1101. [pdf]

Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. and Salakhutdinov, R. (2014)
Dropout: A simple way to prevent neural networks from overfitting
The Journal of Machine Learning Research, 15(1), pp 1929-1958. [pdf]

Srivastava, N., Salakhutdinov, R. R. and Hinton, G. E. (2013)
Modeling Documents with a Deep Boltzmann Machine
arXiv preprint arXiv:1309.6865 [pdf]

Graves, A., Mohamed, A. and Hinton, G. E. (2013)
Speech Recognition with Deep Recurrent Neural Networks
In IEEE International Conference on Acoustic Speech and Signal Processing (ICASSP 2013) Vancouver, 2013. [pdf]

Joseph Turian's map of 2500 English words produced by using t-SNE on the word feature vectors learned by Collobert & Weston, ICML 2008