Guillaume Lajoie

matteo.alleman@mila.quebec

Biography

Guillaume Lajoie is an Associate professor in the Department of Mathematics and Statistics at Université de Montréal and a Core Academic Member of Mila – Quebec Artificial Intelligence Institute. He holds a Canada-CIFAR AI Research Chair, and a Canada Research Chair (CRC) in Neural Computation and Interfacing.

His research is positioned at the intersection of AI and Neuroscience where he develops tools to better understand mechanisms of intelligence common to both biological and artificial systems. His research group's contributions range from advances in multi-scale learning paradigms for large artificial systems, to applications in neurotechnology. Dr. Lajoie is actively involved in responsible AI development efforts, seeking to identify guidelines and best practices for use of AI in research and beyond.

Current Students

Matteo Alleman

Postdoctorate - Université de Montréal

Rohan Banerjee

Collaborating Alumni - Polytechnique Montréal

rohan.banerjee@mila.quebec

Independent visiting researcher - N/A

philippe.beaudoin@mila.quebec

PhD - Université de Montréal

Co-supervisor :

Hugo Larochelle

sangnie.bhardwaj@mila.quebec

Colin Bredenberg

Independent visiting researcher - Université de Montréal

Co-supervisor :

colin.bredenberg@mila.quebec

Leo Choiniere

PhD - Université de Montréal

leo.choiniere@mila.quebec

Eric Elmoznino

PhD - Université de Montréal

Principal supervisor :

Yoshua Bengio

eric.elmoznino@mila.quebec

Leo Gagnon

PhD - Université de Montréal

leo.gagnon@mila.quebec

Tom George

Postdoctorate - McGill University

Principal supervisor :

tomgeorge1@btinternet.com

Juan Guerra

PhD - Polytechnique Montréal

juan.guerra@mila.quebec

thomas.jiralerspong@mila.quebec

Nanda Harishankar Krishna

PhD - Université de Montréal

Anna Jahn

Independent visiting researcher - McGill University

anna.jahn1@mila.quebec

Chen Jiang

PhD - McGill University

Principal supervisor :

Paul Masset

chen.jiang@mila.quebec

Thomas Jiralerspong

PhD - Université de Montréal

Co-supervisor :

Yoshua Bengio

Collaborating researcher - Université de Montréal

Co-supervisor :

Dhanya Sridhar

tejaskasetty@gmail.com

PhD - McGill University

Principal supervisor :

pingsheng.li@mila.quebec

mathys.loiselle@mila.quebec

Mathys Loiselle

Research Intern - Concordia University

Co-supervisor :

Matt Perich

Ximeng Mao

PhD - Université de Montréal

Co-supervisor :

Joelle Pineau

ximeng.mao@mila.quebec

Abdel Mfougouon Njupoun

PhD - Université de Montréal

Co-supervisor :

alexandre.payeur@mila.quebec

abdelnjupoun@gmail.com

PhD - Université de Montréal

Co-supervisor :

Independent visiting researcher - Université de Montréal

Julia Price

Master's Research - Université de Montréal

julia.price@mila.quebec

Mauricio Rivera

Master's Research - Université de Montréal

Principal supervisor :

Marco Bonizzato

mauriv2010@gmail.com

hee-woon.ryoo@mila.quebec

Avery Ryoo

PhD - Université de Montréal

Principal supervisor :

Matt Perich

PhD - Université de Montréal

Co-supervisor :

Lune Bellec

sainathp@mila.quebec

ezekiel.williams@mila.quebec

Ryan Vogt

Postdoctorate - Université de Montréal

ryan.vogt@mila.quebec

Ezekiel Williams

PhD - Université de Montréal

Machine Learning for the Segmentation of Different Nerve Fibre Activations from Brain-to-body Neural Signals

Blog Posts

Représentation graphique d'un nerf vague

May 21, 2025

Param Raval

Olivier Tessier-Larivière

Pascal Fortier-Poisson

Blake Richards

Guillaume Lajoie

Read the article

June 13, 2024

What Do Synaptic Weight Distributions Tell Us About Learning in the Brain ?

Roman Pogodin

Jonathan Cornford

Arna Ghosh

Gauthier Gidel

Guillaume Lajoie

Blake Richards

Read the article

Publications

Defining and Generating Compositional Structure

Eric Elmoznino

Tejas Kasetty

Blaise Agüera y Arcas

Yoshua Bengio

Compositional structure is pervasive in nature and across scales: atoms combine into molecules, molecules into cells, and cells into organis… (see more)ms, while further compositional hierarchies appear within each level. More broadly, compositionality, the reuse and recombination of parts into novel structures, is a fundamental source of interesting complexity in both natural and artificial systems. Yet despite its importance, and despite our apparent ease in recognizing it, mathematical accounts capturing what we intuitively mean by compositional structure remain limited. Here, we fill this gap by formalizing a simple idea: compositional structure emerges when parts, considered together, reveal shared structure that is not captured when those parts are modeled separately. We express this idea using algorithmic information theory, the mathematics of information and compression, which provides a natural way to identify structure in arbitrary data without specifying in advance what that structure must look like. Applied across scales, our definition informs where and how compositional structure emerges within an object. Using tractable approximations, we then invert the framework to generate compositional objects: we specify the desired form of their compositional structure without prescribing their particular content, and optimize directly over strings and images. The resulting objects exhibit readily recognizable hierarchical reuse and part-whole organization, suggesting that the definition captures compositional structure rather than merely complexity, regularity, or shared information. Finally, we discuss implications for machine learning, including a new perspective on depth in neural architectures and principles for constructing compositional data curricula capable of supporting open-ended learning.

2026-10-13

arXiv (preprint)

Gated Target Propagation for Compositional Generalization in Continual Learning

Abdel Mfougouon Njupoun

Colin Bredenberg

Blake Aaron Richards

Continual learning is typically framed as acquiring new knowledge without catastrophically forgetting previous tasks. However, a flexible co… (see more)ntinual learner should also be able to reuse and recombine previously acquired knowledge to rapidly solve novel task compositions. We introduce Gated Target Propagation (GaTaP), a continual learning algorithm in which task-specific gating variables---learned through a closed-form inner loop update---selectively suppress or enhance network modules. Network parameters are learned in a slower timescale outer loop, using the same local difference target propagation error signal as is used for adapting gating variables. We provide tractable experiments on class-incremental learning scenarios for both multilayer perceptron and convolutional network architectures. We show strong performance retention on previously learned tasks, as well as compositional generalization to unseen tasks, achieved through few-shot gain adaptation at inference. We analyze learned gating patterns and find that related tasks exhibit similar gating patterns, suggesting that inferred gates capture meaningful, reusable task structure. Overall, GaTaP provides a powerful framework for jointly ameliorating catastrophic forgetting and enabling few-shot compositional generalization in neural network models.

2026-10-02

arXiv (preprint)

KV-streams for Efficient Compaction in Agentic Reinforcement Learning

Emiliano Penaloza

Dane Malenfant

Dheeraj Vattikonda

Roger Creus Castanyer

Siddarth Venkatraman

Abhay Puri

Jonathan Light

Matthew J. Sargent

Augustine N. Mavor-Parker

M. Caccia

Lucas Caccia

Glen Berseth

Esmeralda S. Whitammer

Minseon Kim

Scaling the horizon of agentic LLMs is bottlenecked by the need to fit ever longer context traces in GPU memory. Context compaction has been… (see more) the most popular mechanism to alleviate this issue, keeping GPU memory constant for a given trace. Unfortunately, most compaction strategies rely on prefilling the LLM context many times over, hindering training throughput. To alleviate this bottleneck and enable efficient trainable compaction, we propose KV-streams, a plug-and-play strategy compatible with any compaction strategy that substantially increases throughput while showing no evidence of hindering performance. KV-streams enable scalable compaction by streaming the KV cache forward rather than flushing it after each compaction. We show that KV-streams enable three different compaction strategies, achieving a 2.6 to 5x wall-clock speedup in training. Beyond efficiency, we find that the streamed KV cache can act as a recurrent state, carrying forward information that has long since disappeared from the context. Specifically, in a controlled setting we show that, contrary to prior work, RL alone is all that is needed for this behavior to emerge. Overall, we show KV-streams to be an efficient and lightweight plug-and-play addition to any post-training pipeline.

2026-09-27

arXiv (preprint)

KV-streams for Efficient Compaction in Agentic Reinforcement Learning

Emiliano Penaloza

Dane Malenfant

Dheeraj Vattikonda

Roger Creus Castanyer

Siddarth Venkatraman

Abhay Puri

Jonathan Light

Matthew J. Sargent

Augustine N. Mavor-Parker

M. Caccia

Lucas Caccia

Glen Berseth

Esmeralda S. Whitammer

Minseon Kim

2026-09-27

arXiv (preprint)

BrainWideBench: Benchmarking large-scale pretraining and across-animal transfer in multi-region neural recordings

Alexandre Andre

Shivashriganesh P. Mahato

Vinam Arora

Keshav Balaji

Divyansha Lachi

Nanda H. Krishna

Jingyun Xiao

Yizi Zhang

Ximeng Mao

Wenrui Ma

Han Yu

International Brain Laboratory

Daniel Birman

Niccolò Bonacchi

Gaelle A. Chapuis

Joana A. Catarino

Felicia Davatolhagh

Mayo Faulkner

Laura Freitas-Silva

Fei Hu … (see 22 more)

Julia M. Huntenburg

Anup Khanal

Inês Laranjeira

Petrina Lau

Guido T. Meijer

Nathaniel J. Miska

Jean-Paul Noel

Alejandro Pan-Vazquez

Georg Raiser

Cyrille Rossant

Karolina Z. Socha

Anne E. Urai

Miles J. Wells

Steven J. West

Olivier Winter

Cole Hurwitz

Mehdi Azabou

Matthew R. Whiteway

Liam Paninski

Eva L. Dyer

Advances in large-scale neural recording have made it possible to collect data across many animals and distributed brain regions, raising th… (see more)e question of whether this scale can be exploited to learn general-purpose neural representations transferable across diverse downstream tasks. Yet, progress toward this goal has been limited by fragmented evaluation protocols and a narrow focus on individual task domains. Here, we present BrainWideBench, a benchmark for evaluating across-animal transfer on multi-region neural recordings, built on the International Brain Laboratory Brainwide Map dataset of neural and behavioral recordings spanning 276 brain regions from 139 mice performing a sensory-guided decision-making task. The benchmark is organized around three complementary task suites that evaluate whether learned representations support downstream decoding of behavior, can predict masked or future neural activity, and can recover biologically meaningful anatomical organization. With this benchmark, we systematically evaluate pretraining methods across transfer settings, including finetuning on downstream objectives and zero-shot generalization to unseen animals. Our results confirm pretraining improves performance over matched single-session baselines, but we show current methods exhibit heterogeneity in transfer capabilities: gains depend strongly on the alignment between pretraining objectives and downstream tasks. No single approach performs uniformly well across all three suites, and most methods are designed to only address a subset of them. Together, these findings suggest that learning representations that jointly generalize across behavior, dynamics, and anatomy remains an open challenge. By providing a unified and reproducible evaluation suite, BrainWideBench establishes a framework for measuring progress toward general-purpose models of the mouse brain.

2026-09-17

arXiv (preprint)

Assistive algorithms influence neural representations in motor brain-computer interfaces

Pavithra Rajeswaran

Alexandre Payeur

Amy L. Orsborn

Task errors are used to learn and refine motor skills. We investigated how task assistance influences learned neural representations using B… (see more)rain-Computer Interfaces (BCIs), which map neural activity into movement via a decoder. We analyzed motor cortex activity as monkeys practiced BCI with a decoder that adapted to improve or maintain performance over days. Over time, task-relevant information became concentrated in fewer neurons, unlike with fixed decoders. At the population level, task information also became largely confined to a few neural modes that accounted for a small fraction of the population variance. A neural network model suggests the adaptive decoders directly contribute to forming these more compact neural representations. Our findings suggest that assistive decoders manipulate error information used for long-term learning computations like credit assignment, which may explain the altered neural representations and inform real-world BCI design. Assistive algorithms are widely used in brain-computer interfaces (BCIs), but their effects on neural representations remain unclear. Here, the authors show that adaptive BCIs lead the brain to learn compact representations, suggesting assistive algorithms shape long-term credit assignment.

2026-09-14

Nature Communications (published)

Exponentiated gradient learning yields brain-like synaptic distributions

Jonathan H. Cornford

Roman Pogodin

Arna Ghosh

Kaiwen Sheng

Brendan A. Bicknell

Olivier Codol

Pingsheng Li

Beverley A. Clark

Blake A. Richards

Artificial neural networks trained with gradient descent (GD) are widely used to model brain learning, but GD allows synapses to switch betw… (see more)een excitatory and inhibitory and produces weight distributions unlike those observed experimentally. We show that an alternative rule, exponentiated gradient (EG), avoids these problems because its multiplicative, exponential weight updates mathematically prohibit synapses from changing sign and preferentially scale already-large weights, naturally generating log-normal distributions. This mechanism also concentrates task-relevant computation onto a smaller subset of strong connections, which we show makes EG-trained recurrent neural networks more robust to synaptic pruning and better able to learn from sparsely relevant inputs than GD-trained networks, while matching GD's performance on standard cognitive tasks. These findings tie a normative optimization principle directly to observed features of synaptic organization in the brain.

2026-08-31

Neuron (published)

A game theory for foundation models shows new paths to rational cooperation through similarity inference

Alexander Meulemans

Maciej Wołczyk

Marissa A. Weis

Rajai Nasser

Roberta Rocca

Seijin Kobayashi

Angelika Steger

Marcus Hutter

James Manyika

Rif A. Saurous

João Sacramento

Blaise Agüera y Arcas

As autonomous agents powered by foundation models are increasingly integrated into social and economic systems, understanding the principles… (see more) governing their collective behavior is essential for ensuring safety and cooperation. Classical game theory, the dominant framework for modeling rational interaction, is built upon the assumption of `decoupled agency,'where agents treat their own decision-making as independent of the environment and other actors. Modern AI agents, however, jointly predict their own future actions alongside external observations. Here, we report a striking finding: when interacting in stylized social dilemmas, foundation model agents engaging in optimal planning consistently converge to stable cooperation, directly contradicting classical game-theoretic predictions of mutual defection. To understand this phenomenon, we introduce the `embedded Bayesian agent,'a theoretical model for foundation model agents. By shifting from decoupled to embedded agency, these agents model themselves as part of the universe they inhabit, maintaining epistemic uncertainty about their own decision-making algorithms. We show that by inferring whether others are behaviorally similar, an embedded agent treats its own deliberation during planning as evidence: a decision to cooperate predicts a similar decision by a similar partner. We formalize this mechanism of similarity inference through the `embedded equilibrium,'a novel solution concept replacing the Nash equilibrium to provide a foundational game theory for the social behavior of modern AI agents.

2026-08-03

arXiv (preprint)

Recursive Vision Language Models for General Symbolic Reasoning

Omid Nejati Manzari

Hassan Rivaz

Hard symbolic-reasoning tasks such as Sudoku, maze pathfinding, and ARC remain challenging for LLMs due to their fixed-depth autoregressive … (see more)reasoning, which limits systematic search, refinement, and backtracking. While recursive models such as Hierarchical Reasoning Model (HRM) and Tiny Recursive Model (TRM) address this limitation through iterative latent-state refinement, they are typically task-specific and do not leverage pretrained language priors. We propose R-Qwen, a recursive reasoning framework built upon a pretrained Qwen backbone. R-Qwen repeatedly refines a candidate solution through programmatic self-recursion and deep supervision, combining the structured iterative computation of recursive models with the linguistic and reasoning priors of pretrained LLMs. We further adapt Hierarchical Supervision Weighting (HSW) to autoregressive models by exponentially weighting losses across recursive steps. HSW reduces gradient variance by at least 50\%, improves the signal-to-noise ratio of stochastic gradients, and accelerates convergence. Across eight challenging benchmarks, R-Qwen consistently outperforms prior recursive reasoning models and substantially larger LLMs while using a comparable number of trainable parameters. Notably, on ARC-AGI dataset, our model achieves a 27.6\% improvement over the baseline, highlighting the effectiveness of recursive refinement for general symbolic reasoning. These results suggest that recursive reasoning mechanisms and pretrained language model priors are complementary approaches for improving symbolic puzzle-solving. Code and models will be released after acceptance.

2026-08-01

arXiv (preprint)

Leveraging unlabelled data for generalizable neural population decoding

Ximeng Mao

Nanda H. Krishna

Avery Hee-Woon Ryoo

Matthew G. Perich

Robust and accurate neural decoders are integral to neurotechnologies such as brain-computer interfaces and closed-loop experiments. Recent … (see more)work has shown that tokenizing neural data at the spike level facilitates multi-session pretraining and delivers state-of-the-art decoding performance. However, current spike-based models are restricted to supervised learning (SL), limiting training to datasets with paired behavioural labels. To address this limitation, we introduce MOJO (Masked autOencoder-based JOint training), a training framework for spike-tokenizing models that jointly leverages self-supervised learning (SSL) via masked autoencoding and SL objectives. We evaluate MOJO on three spiking datasets spanning monkey motor cortex during reaching tasks and multi-regional mouse recordings during vision and decision making tasks, demonstrating superior performance over purely SL-trained models. This improvement is especially pronounced when training with limited labelled data, particularly in few-shot finetuning, where only a small amount of labelled data from a new session is available. Incorporating SSL also yields more interpretable neuronal representations, improving performance on brain region classification and spike-statistics prediction without explicit optimization for these tasks. We further show that MOJO generalizes beyond spiking data to human electrocorticography during speech, where it continues to outperform purely SL-trained models and achieves performance comparable to neuro-foundation models (NFMs) designed specifically for continuous signals. Overall, augmenting spike-tokenizing models with SSL improves performance in label-impoverished settings and enables the use of unlabelled data across various tasks and species, while generalizing to other neural modalities. These results suggest a path towards more flexible and scalable data usage when training NFMs.

2026-07-14

arXiv (preprint)

Can In-Context Learning Support Intrinsic Curiosity?

Eric Elmoznino

Sangnie Bhardwaj

Johannes Von Oswald

Rajai Nasser

Blaise Agüera y Arcas

João Sacramento

Rif A. Saurous

Effective machine learning depends not only on how we model data, but also on what data we choose to collect. While large sequence models ha… (see more)ve revolutionized data modeling, the problem of automated data selection, or"intrinsic curiosity", remains a significant challenge. Classic approaches incentivize exploration by rewarding an agent based on its"learning progress", which measures how much a newly acquired observation improves a world model's predictive ability. However, evaluating these rewards traditionally requires expensive inner loops of gradient descent updates within each trajectory, rendering them computationally impractical at scale. In this work, we investigate whether the emergent in-context learning (ICL) capabilities of sequence models can eliminate this bottleneck by serving as immediate, update-free world models. Specifically, we evaluate whether an exploration policy can be trained to maximize learning progress, using solely the prediction errors and counterfactual context manipulations of an in-context learner. We first prove that in general Markov decision processes, this is in fact impossible in an unbiased way: the resulting intrinsic rewards either suffer from nuisance terms that bias their estimation of true learning progress, or they cannot be implemented using an in-context learner's prediction errors. Conversely, we prove a positive result for a broad subclass of non-temporal settings, encompassing active learning and Bayesian Experimental Design: here, ICL-derived rewards successfully bound and asymptotically converge to the true learning progress. We corroborate our theory with controlled experiments across continuous and symbolic environments, demonstrating that our ICL-driven framework successfully trains curious data-collection policies that explore optimally.

2026-06-16

arXiv (preprint)

Simplifying the Modeling of Arbitrary Conditionals in Natural Language

Causal Transformers model sequences through an autoregressive factorization of the joint distribution, which enables efficient left-to-right… (see more) decoding and conditional likelihood computation. However, they cannot tractably sample from or evaluate arbitrary conditionals -- e.g., a block of text conditioned on past and future tokens. Recent work aims to solve this problem through novel architectures, but they often lead to sub-optimal modeling of such conditionals and degraded generations. We propose Arbitrary Conditionals GPT (AC-GPT) which introduces a simple modification to standard causal Transformers to enable evaluating and sampling from arbitrary conditionals -- including past, future, and mixed contexts -- within a single forward pass. Unlike prior approaches, our method preserves the standard left-to-right ordering and next-token prediction objective essential for both strong performance and efficient training on natural language. Crucially, this compatibility allows existing LLMs to be fine-tuned for arbitrary conditioning. Our empirical results indicate that our method outperforms baselines on modeling arbitrary conditionals, without degrading standard left-to-right performance.

2026-06-11

arXiv (preprint)