SAIL Montreal is happy to have a strong presence at ICML 2023. Our office manager Geneviève Bernard will be present on site along with some of our research scientist who are happy to have a chat with you.

What is SAIL Montreal?

SAIL Montreal (SAIT AI Lab Montreal) is a recently-established academic-style research lab (in close collaboration with Mila) from Samsung whose mission is to push the boundaries of AI with SAIT and Mila through open fundamental research. Headed by Simon Lacoste-Julien, professor in computer science at Université de Montréal and co-founding member of Mila, SAIL is located in Mila's corporate space at the heart of the Montreal AI ecosystem nearby Borealis AI, Meta AI, Microsoft Research and others where an open collaborative environment is encouraged.

Accepted papers at ICML 2023

Click the links for more information on these papers (requires ICML registration). Underlined authors are SAIL-affiliated, * denotes equal contribution between authors.

Oral

• Difference of submodular minimization via DC programming [openreview] [code]
  Marwa El Halabi, George Orfanides, Tim Hoheisel
  Oral presentation: Oral C1 supervised learning Thu 27 Jul 3 p.m. HST — 4:30 p.m. HST
  Exhibit Hall 1 #643, Wed 26 Jul 11 a.m. HST — 12:30 p.m. HST

Tuesday

• Unlocking Slot Attention by Changing Optimal Transport Costs [openreview] [code]
  Yan Zhang*, David W Zhang*, Simon Lacoste-Julien, Gertjan J. Burghouts, Cees G. M. Snoek
  Exhibit Hall 1 #111, Tue 25 Jul 11 a.m. HST — 1:30 p.m. HST

• Can We Scale Transformers to Predict Parameters of Diverse ImageNet Models? [openreview] [code]
  Boris Knyazev, Doha Hwang, Simon Lacoste-Julien
  Exhibit Hall 1 #808, Wed 26 Jul 11 a.m. HST — 12:30 p.m. HST

Wednesday

• Fairness in Streaming Submodular Maximization over a Matroid Constraint [openreview] [code]
  Marwa El Halabi*, Ashkan Norouzi-Fard*, Jakab Tardos*, Jakub Tarnawski*, Federico Fusco*
  Exhibit Hall 1 #642, Wed 26 Jul 11 a.m. HST — 12:30 p.m. HST

• Difference of submodular minimization via DC programming [openreview] [code]
  Marwa El Halabi, George Orfanides, Tim Hoheisel
  Oral presentation: Oral C1 supervised learning Thu 27 Jul 3 p.m. HST — 4:30 p.m. HST
  Exhibit Hall 1 #643, Wed 26 Jul 11 a.m. HST — 12:30 p.m. HST

• CrossSplit: Mitigating Label Noise Memorization through Data Splitting [openreview]
  Jihye Kim, Aristide Baratin, Yan Zhang, Simon Lacoste-Julien
  Exhibit Hall 1 #210, Wed 26 Jul 11 a.m. HST — 12:30 p.m. HST

• Fast Online Node Labelling with Local and Global Consistency [openreview]
  Baojian Zhou, Yifan Sun, Reza Babanezhad
  Exhibit Hall 1 #723, Wed 26 Jul 2 p.m. HST — 3:30 p.m. HST

• Target-based Surrogates for Stochastic Optimization [openreview]
  Jonathon W. Lavington, Sharan Vaswani, Reza Babanezhad, Mark Schmidt, Nicolas Le Roux
  Exhibit Hall 1 #438, Wed 26 Jul 2 p.m. HST — 3:30 p.m. HST

Thursday

• Equivariance with Learned Canonicalization Functions [openreview]
  Sékou-Oumar Kaba*, Arnab Kumar Mondal*, Yan Zhang, Yoshua Bengio, Siamak Ravanbakhsh
  Exhibit Hall 1 #127, Thu 27 Jul 1:30 p.m. HST — 3 p.m. HST

Finding us

The following research scientists will be available in-person at ICML 2023:

• Boris Knyazev
• Reza Babanezhad
• Marwa El Halabi
• Yash Goyal
• Yan Zhang
• Jihye Kim (Visiting Research Scientist)

Come find us at our posters, or send us an email to arrange something individual. You can find the full list of people on our main website.

Workshops

We also have some workshop papers:

Friday

• Pretrained Language Models to Solve Graph Tasks in Natural Language
  Frederik Wenkel, Guy Wolf, Boris Knyazev
  ICML Workshop on Structured Probabilistic Inference & Generative Modeling

• Learning to Optimize with Recurrent Hierarchical Transformers
  Abhinav Moudgil, Boris Knyazev, Guillaume Lajoie, Eugene Belilovsky
  ICML Workshop on New Frontiers in Learning, Control, and Dynamical Systems

• Neural Networks Are Graphs! Graph Neural Networks for Equivariant Processing of Neural Networks
  David W. Zhang, Miltiadis Kofinas, Yan Zhang, Yunlu Chen, Gertjan J. Burghouts, Cees G. M. Snoek
  ICML Workshop on Topology, Algebra, and Geometry in Data Science

Saturday

• Decision-Aware Actor-Critic with Function Approximation and Theoretical Guarantees
  Sharan Vaswani, Amirreza Kazemi, Reza Babanezhad, Nicolas Le Roux
  ICML Workshop on Duality Principles for Modern ML

SAIL Montreal is happy to sponsor NeurIPS 2022! We will have a booth where you can meet some of our research scientists.

Continue Reading