DeepTutorial2009.pptTutorial: Learning Deep Architectures Yoshua Bengio, U. Montreal Yann LeCun, NYU ICML Workshop on Learning Feature Hierarchies, June 18th, 2009, Montreal Deep Motivations Brains have a deep architecture Humans organize their ideas hierarchically, through composition of simpler ideas Unsufficiently deep architectures can be exponentially inefficient Distributed (possibly sparse) representat
Deep Learning II Russ Salakhutdinov Department of Statistics and Computer Science! University of Toronto h7p://www.utstat.toronto.edu/~rsalakhu/isbi.html Talk Roadmap • Deep Boltzmann Machi
