Benjamin Scellier

I’m a Ph.D. candidate at MILA supervised by Yoshua Bengio.

My interests include Deep Learning, Unsupervised Learning, Generative Models, Recurrent Neural Networks, Attention Mechanisms…

I have been doing research at the interface between Deep Learning and Computational Neuroscience. The question is whether the brain could implement a learning algorithm similar to Backpropagation. This led to a new training framework for energy based models that we have called Equilibrium Propagation.