Title: Deep Learning
Speaker: Laurent Charlin (HEC Montreal), Mike Pieper (MILA, Montreal Institute for Learning Algorithms, University of Montreal), Arsène Tchango (MILA) and Margaux Luck (MILA)

Abstract: This workshop will introduce the audience to the field of deep learning. We will first focus on defining this field in comparison to the field of statistics pointing out differences both in aim and in methodology. We will then introduce the fundamental concepts that underlie modern neural network models. In the afternoon, we will focus on two applications that contributed greatly to the field's recent notoriety: image modelling and text modelling. Convolutional neural networks (CNNs) and recurrent neural networks (RNNs) will be discussed in that context.

Tentative schedule: 

• Statistics and Deep Learning (90 min) -- Laurent Charlin
• Introduction to deep learning (90 min) -- Mike Pieper
• Deep learning for text (90 min) -- Arsène Tchango
• Deep learning for images (90 min) -- Margaux Luck and Tristan Sylvain