The Montreal AI Symposium aims at gathering experts and professionals interested in fundamental advances and applications of artificial intelligence, with an emphasis on machine learning, deep learning and related approaches.
The Symposium welcomes both academic and industrial participants; it seeks to build strong connections between researchers within the Greater Montreal area.

We will feature a day-long event, filled with keynote addresses, contributed talks and posters, and time for networking and socializing.



The Symposium will be held on August 28th, 2018, and is hosted in the Pollack Hall of the Wirth Music Building of McGill University. The event is free of charge for participants with mandatory registration. It is easily accessible from the public transit system (“McGill” metro station – Green line).

schulich hall_en


Daycare will be offered for free during the symposium. However, registration will be mandatory.

Code of conduct

This symposium aims at facilitating discussions and the exchange of ideas. We ask all participants to be respectful of others and to be aware of their own behaviour. Registration for the symposium will require signing the code of conduct available here.

Accepted Papers

The list of accepted papers is available here.


8.00 – 9.00 Registration
9.00 – 9.10 Opening Remarks
9.10 – 9.50 Keynote – Margaret Mitchell
9.50 – 10.10 Duckietown: a Platform for Teaching, Robotics and Machine Learning Research
Maxime Chevalier-Boisvert (MILA); Manfred R Diaz (MILA); Breandan M Considine (MILA)
10.10 – 10.30 Using Reward Machines for High-Level Task Specification and Decomposition in Reinforcement Learning
Richard Valenzano (Element AI); Sheila A. McIlraith (University of Toronto); Rodrigo A Toro Icarte (University of Toronto); Toryn Klassen (University of Toronto)
10.30 – 11.00 Coffee break
11.00 – 11.20 Parametric Divergences are Good Task Losses for Generative Modeling
Gabriel Huang (MILA); Hugo Berard (MILA); Ahmed Touati (MILA); Gauthier Gidel (MILA); Pascal Vincent (Université de Montréal); Simon Lacoste-Julien (University of Montreal)
11.20 – 11.40 A Variational Inequality Perspective on Generative Adversarial Networks
Gauthier Gidel (MILA); Hugo Berard (MILA); Gaëtan Vignoud (MILA); Pascal Vincent (University of Montreal); Simon Lacoste-Julien (University of Montreal)
11.40 – 12.00 Learning the Base Distribution in Implicit Generative Models
Cem Subakan (MILA); Sanmi Koyejo (University of Illinois at Urbana-Champaign); Paris Smaragdis (University of Illinois at Urbana-Champaign)
12.00 – 13.30 Lunch
13.30 – 14.10 Keynote – Melanie Mitchell
14.10 – 14.30 Fashion-Gen: The Generative Fashion Dataset and Challenge
Negar Rostamzadeh (Element AI); Seyedarian Hosseini (Element AI); Thomas Boquet (Element AI); Wojciech Stokowiec (Element AI); Ying Zhang (Element AI); Christian Jauvin (Element AI); Chris Pal (Element AI)
14.30 – 14.50 Learning to Learn with Conditional Class Dependencies
Xiang Jiang (Imagia Cybernetics, Dalhousie University); Mohammad Havaei (Imagia Cybernetics)
14.50 – 15.20 Coffee Break
15.20 – 15.40 Deep Learning Method for Smart Charging of Electric Vehicles
Marc-André Gardner (Université Laval); Karol Lopez (Université Laval); Christian Gagné (Université Laval)
15.40 – 16.00 Transfer Deep Reinforcement Learning for Residential Energy Management
Di Wu (McGill University); guillaume rabusseau (McGill); Vincent Francois-Lavet (McGill University); Doina Precup (McGill University); Benoit Boulet (McGill University)
16.00 – 16.20 Deep Photovoltaic Nowcasting
Jinsong Zhang (Université Laval); Jean-Francois Lalonde (Université Laval)
16:20 – 17.00 Panel
17.00 – 20.00 Poster Session + Happy Hour with Sponsors


Melanie Mitchell

Melanie Mitchell is Professor of Computer Science at Portland State University, and External Professor at the Santa Fe Institute. She attended Brown University, where she majored in mathematics, and the University of Michigan, where she received a Ph.D. in computer science.   Her dissertation, in collaboration with her advisor Douglas Hofstadter, was the development of “Copycat”, a computer program that makes human-like analogies in an idealized domain.  Her current work is on “Situate”, an extension of Copycat that interprets and makes analogies between real-world visual situations.  Melanie is the author or editor of five books and numerous scholarly papers in the fields of artificial intelligence, cognitive science, and complex systems. Her most recent book, Complexity: A Guided Tour (Oxford, 2009), won the 2010 Phi Beta Kappa Science Book Award. It was also named by as one of the ten best science books of 2009, and was longlisted for the Royal Society’s 2010 book prize.   She is currently at work on a new book about the current state of artificial intelligence and the prospects for human-level AI.

Keynote title: Artificial Intelligence and the “Barrier of Meaning”

In 1986, the mathematician and philosopher Gian-Carlo Rota wrote, “I wonder whether or when artificial intelligence will ever crash the barrier of meaning.” Here, the phrase “barrier of meaning” refers to a belief about humans versus machines: humans are able to “actually understand” the situations they encounter, whereas AI systems (at least current ones) do not possess such understanding. The internal representations learned by (or programmed into) AI systems do not capture the rich “meanings” that humans bring to bear in perception, language, and reasoning.
In this talk I will assess the state of the art of artificial intelligence in several domains, and describe some of the current limitations and vulnerabilities, which can be accounted for by a lack of true understanding of the domains these systems work in. I will explore the following questions: (1) To be reliable in human domains, what do AI systems actually need to “understand”? (2) Which domains require human-like understanding? And (3) What does such understanding entail?


Margaret Mitchell

Margaret Mitchell is a Senior Research Scientist in Google’s Research & Machine Intelligence group, working on artificial intelligence, multimodality, and ethics. Her research involves vision, language, computer vision, and grounded language generation, focusing on how to evolve artificial intelligence towards positive goals. This includes research on helping computers to communicate based on what they can process, as well as projects to create assistive and clinical technology from the state of the art in AI.

Keynote title: To be announced



Before the poster session, there will be a panel on the Montreal AI ecosystem. The panelists will be:

  • Yoshua Bengio, professor at Université de Montréal
  • Narjès Boufaden, founder and CEO of Keatext
  • Sylvain Carle, partner at Real Ventures
  • Mark Maclean, senior director of Montreal International
  • Joelle Pineau, professor at McGill University and head of Facebook Montreal


The event is free of charge for participants. However, all attendees must register to attend the symposium. To prevent the event from being full immediately, we will open an unlimited number of registrations until August, 8th.

Click here to access the registration form

At the end of that month, participants will be chosen at random amongt those who registered, based on the capacity of the venue. For every accepted paper, one of its authors will have a guaranteed participation. The random draw will not be uniform as the organizers will try to get as representative an attendance as possible. To that end, we encourage you to fill in the demographics questionnaire which, with your approval, will be used for the random draw.

Click here to fill the demographics questionnaire

Once the participations have been chosen, each participant will need to confirm their attendance. If they do not, the registration will be relinquished. Participants who cannot attend can also choose to voluntarily relinquish their registration, and this until 2 days prior to the event. Participants who are registered but do not come to the symposium might be barred from next year’s event. This is to ensure as many people as possible will attend this event.

Call for Contributions

We invite you to submit a contribution to the second Montreal AI symposium. We encourage submission of abstracts from both academic and industrial researchers, which can describe a technical or practical contribution, an open problem, an application or a position statement. Previously published material is acceptable. Do not include any confidential material.

Topics of interest include but are not limited to:

  • Fundamental research in deep learning, reinforcement learning, kernel machines, Bayesian modeling, ensemble methods, optimization for machine learning;
  • Implementation issues, parallelization, software platforms, hardware;
  • Applications, including vision, audio, speech, natural language processing, robotics, healthcare, bioinformatics.

Instructions for abstract submission

Submissions will be handled electronically via the symposium’s CMT website: You will be asked to create a CMT account if you do not already have one. All abstracts must be written in English.

To create a new submission, click on ‘Create new submission’. You will be asked to select your subject area and indicate your conflicts of interest as well as your preference for presentation format (short talk or poster).  Abstracts should be at most two pages in 10pt font, not including references.

Reviews are double-blind: submissions revealing the authors’ identities will be automatically rejected. When citing your previous work, refrain from using ‘we’ or ‘our’. We accept previously published material (including pre-prints) but we ask you not to cite this material to maintain anonymity during the reviewing period. We also ask you to omit acknowledgments in your submission.

Submission deadline: July 13 2018, 17:00 EST.

Notification of acceptance: Early August 2018

Call for Sponsors

The organization of this event is entirely supported by sponsors, as registration is free of charge for attendees.

If you are interested in learning about our sponsorship offers for the Montreal AI Symposium, please contact the Symposium organizers.

Territorial acknowledgment

McGill is situated on land which has long served as a site of meeting and exchange amongst Indigenous peoples, including the Haudenosaunee and Anishinabeg nations. The Montreal AI Symposium honours and respects the diverse Indigenous peoples connected to this territory on which we shall gather.

Organizing Committee

Senior Program Chair: Layla El Asri, Microsoft Research Montréal

Program Chairs:

  • Philippe Beaudoin, Element AI
  • Simon Lacoste-Julien, MILA, Université de Montréal
  • Nicolas Le Roux, Google Brain and McGill University

Local Chair: Jessica Mastronardi, Microsoft Research Montréal

Inclusion and diversity chairs:

  • Jessica Thompson, MILA
  • Laurent Dinh, Google Brain

Contact the organizers:

Contact the inclusion and diversity chairs:


Platinum Sponsors



Gold Sponsors

google       deepmind                     coveo           Behavox_logo

element_ai            imagia

ubisoft             wavo-logo-black      Mila_LOGO_HORIZONTALE_standard_k

Silver Sponsors

CREATOR: gd-jpeg v1.0 (using IJG JPEG v62), quality = 82        huawei               logo_samsung                 automat       kinova      morgan_stanley   logo-CRIM