Since the 90s we have known about the fundamental challenge in training a parametrized dynamical system such as a recurrent networks to capture long-term dependencies. The notion of stable memory is crucial in understanding this issue, and is behind the LSTM and GRU architectures, as well as the recent work on networks with an external memory. We present several new ideas exploring how to further expand the reach of recurrent architectures, improve their training and scale up their memory, in particular to model language-related data and better capture semantics for question answering, machine translation and dialogue.
Yoshua Bengio received a PhD in Computer Science from McGill University, Canada in 1991. After two post-doctoral years, one at M.I.T. with Michael Jordan and one at AT&T Bell Laboratories with Yann LeCun and Vladimir Vapnik, he became professor at the Department of Computer Science and Operations Research at Université de Montréal. He is the author of three books and more than 200 publications, the most cited being in the areas of deep learning, recurrent neural networks, probabilistic learning algorithms, natural language processing and manifold learning. He is among the most cited Canadian computer scientists and is or has been associate editor of the top journals in machine learning and neural networks. Since '2000 he holds a Canada Research Chair in Statistical Learning Algorithms, is a Senior Fellow of the Canadian Institute for Advanced Research and since 2014 he co-directs its program focused on deep learning. He heads the Montreal Institute for Learning Algorithms (MILA), currently the largest academic research group on deep learning. He is on the board of the NIPS foundation and has been program chair and general chair for NIPS. He has co-organized the Learning Workshop for 14 years and co-created the new International Conference on Learning Representations. His current interests are centered around a quest for AI through machine learning, and include fundamental questions on deep learning and representation learning, the geometry of generalization in high-dimensional spaces, generative models, biologically inspired learning algorithms, natural language understanding and other challenging applications of machine learning.