|Stochastic Systems Group|
An Efficient Way to Learn Deep Generative Models
Canadian Institute for Advanced Research and University of Toronto
I will describe an efficient, unsupervised learning procedure for deep generative models that contain millions of parameters and many layers of hidden features. The features are learned one layer at a time without any information about the final goal of the system. After the layer-by-layer learning, a subsequent fine-tuning process can be used to significantly improve the generative or discriminative performance of the multilayer network by making very slight changes to the features.
This approach leads to excellent generative models of handwritten digits and natural image patches. It can also be used to create hash functions that map similar documents to similar addresses, thus allowing hash functions to be used for retrieving similar documents in a time that is independent of the size of the database.
Problems with this site should be emailed to firstname.lastname@example.org