Neuro-AI Computations Research Group
biological and artificial neural dynamics
Guillaume Lajoie (PI)
Version Française en construction !
About Guillaume Lajoie
Assistant Professor
Mila - Québec AI Institute (Core Academic Member)
Research Scholar, Fond de Recherche du Québec en Santé (FRQS)
Research center affiliations:
Departmental Accreditations (UdeM)
Contact & Links
email: g.lajoie 'at' umontreal.ca
snail mail: Dept. de Mathématiques et Statistiques, Université de Montréal, 2920 chemin de la Tour, Montréal, QC, H3T 1J4
some code: github.com/glajoie/
CV: link
this site's sections: people / publications / teaching
I am an applied mathematician interested in the interactions and commonalities of biological and artificial neural computations. My research group works at the intersection of AI and Neuroscience, or Neuro-AI, developing tools to better understand neural networks as well as algorithms for brain-machine interfaces for scientific and clinical use. My work is motivated by the remarkable ability of neural networks (biological and artificial) to learn and support complex, emergent computations. I use tools from dynamical systems, information theory, statistics and machine learning to address a range of problems, in collaboration with experimental neuroscientists and machine intelligence researchers.
News
06/2022
Paper on opportunity cost of time and urgency signals in the brain is out in PloS Comp Biol.! Amazing work lead by Max Puelma Touzel
01/2022: Four abstracts from the lab accepted at COSYNE22
Top-down optimization recovers biological coding principles of single-neuron adaptation in RNNs, V. Geadah, G. Kerg, S. Horoi, G. Wolf, G. Lajoie
Deliberation gated by opportunity cost adapts to context with urgency in non-human primates, M. Puelma Touzel, P. Cisek, G. Lajoie
Dissecting emergent network noise compensation mechanisms in working mem- ory tasks, C. Bredenberg, M. Puelma Touzel, R. Engelken, G. Lajoie
Beyond accuracy: robustness and generalization properties of biologically plausible learning rules, Y. H. Liu, G. Lajoie
01/2022: Two papers from the lab accepted at ICLR2022, both with spotlight talks!!!
02/2021 Huge congratulations to lab members Giancarlo Kerg, Ezekiel Williams and Sarthak Mittal for their UNIQUE Center Excellence Scholarships !
01/2021: 5 submissions from the lab accepted at COSYNE 2021 !
PNS-GAN: conditional generation of peripheral nerve signals in the wavelet domain via adversarial networks, Olivier Tessier-Lariviere, Luke Prince, Pascal Fortier-Poisson, Lorenz Wernisch, Emil Hewage, Oli Armitage, Blake Richards*, Guillaume Lajoie*
Urgency as the neural correlate of the opportunity cost of time commitment in de- cision making tasks, Maximilian Puelma Touzel, Paul Cisek, Guillaume Lajoie
Noisy gradient updates drive dimensionality compression in neural networks, Matthew Farrell, Stefano Recanatesi, Madhu Advani, Guillaume Lajoie, Eric Shea-Brown
Learning Function From Structure In Neuromorphic Networks, Laura Suarez, Blake Richards, Guillaume Lajoie, Bratislav Misic
Network-level computational advantages of single-neuron adaptation, Victor Geadah, Giancarlo Kerg, Stefan Horoi, Guy Wolf, Guillaume Lajoie
Keep you eyes out for these !!!!
09/2020: Short opinion piece about the future of automated neurostimualtion optimization , now live on IEEE Brain BrainInsight -- Bayesian optimization for automated neurostimulation: future directions and challenges
09/2020: New paper accepted at NeurIPS2020! See preprint here: Untangling tradeoffs between recurrence and self-attention in neural networks
07/2020: Sarthak's paper accepted at ICML! Learning to Combine Top-Down and Bottom-Up Signals in Recurrent Neural Networks with Attention over Modules
06/2020: Sam's paper published in IEEE TNSRE! Hierarchical Bayesian Optimization of Spatiotemporal Neurostimulations for Targeted Motor Outputs
Funding
NSERC / CRSNG
FRQNT
FRQS
IVADO
Medteq
Mitacs
Google Research
