CERC-AAI Lab
Canada Excellence Research Chair in Autonomous AI
AGI Collective Discord
AI Foundation Models R&D
Canada Excellence Research Chair (CERC) in Autonomous Artificial Intelligence Lab at the Department of Computer Science and Operations Research of the University of Montreal (UdeM) is led by Irina Rish, a full professor at UdeM and Mila - Quebec AI Institute. The CERC-AAI program is supported by the CERC grant from the Canadian government, which is a part of the comprehensive strategy to encourage the development and deployment of artificial intelligence in Canada.
The high-level objective of this CERC in Autonomous AI program is to develop a set of sustained initiatives that will advance AI towards more general, human-like intelligence, by making AI more flexible, robust and broad. More specifically, the lab's research is focused on:
Building large-scale AI foundation models on supercomputers
Neural scaling laws and emergent behaviors: understanding and predicting scaling properties, including emergence, in such models, with respect to both capabilities and alignment metrics.
Continual Learning, Robustness, and OoD Generalization
Nonlinear dynamical neural systems
RL and bio-inspired models of intelligent behavior
With this CERC Chair, Dr. Irina Rish proposes to augment AI by developing better computational models and connecting AI research with vast amounts of knowledge about the mind and brain from biology, neuroscience, psychology, philosophy, and other disciplines. This CERC program also aims at building a community of researchers who can bridge the gap between these disciplines, paving the way toward a new, unified science of both biological and artificial intelligence.
Recent documents, presentations, and podcasts outlining the key philosophy behind this CERC-AAI program:
July 2023: Overview of CERC Research Activities
Continual Learning at Scale - Irina Rish -CoLLAs 2023
AGI, Complex Systems, Transhumanism (NeurIPS 2022)
Build artificial general intelligence (Dec 2020)
Irina Rish—AGI, Scaling, Alignment (Oct 2022)
Out-of-distribution generalization (March 2022)
Autonomous AI (2020)