Sarath Chandar

Biographie

Sarath Chandar est professeur associé au départment de génie informatique et génie logiciel de Polytechnique Montréal, où il dirige le laboratoire de recherche Chandar. Il est également membre académique principal à Mila – Institut québécois d’intelligence artificielle, et titulaire d'une chaire en IA Canada-CIFAR et d'une Chaire de recherche du Canada en apprentissage machine permanent.

Ses recherches portent sur l'apprentissage tout au long de la vie, l'apprentissage profond, l'optimisation, l'apprentissage par renforcement et le traitement du langage naturel. Pour promouvoir la recherche sur l'apprentissage tout au long de la vie, Sarath Chandar a créé la Conférence sur les agents d'apprentissage tout au long de la vie (CoLLAs) en 2022 et a présidé le programme en 2022 et en 2023. Il est titulaire d'un doctorat de l'Université de Montréal et d'une maîtrise en recherche de l'Indian Institute of Technology Madras.

Étudiants actuels

Ista Abbes

Maîtrise recherche - UdeM

Alex Aselstyne

Maîtrise recherche - Polytechnique

Davide Baldelli

Doctorat - Polytechnique

Co-superviseur⋅e :

joe Ben

Stagiaire de recherche - Polytechnique

joumenbensaid@gmail.com

Milan Bhan

Collaborateur·rice de recherche

Diego Cerda Mardini

Maîtrise recherche - McGill

Antoine Clavaud

Maîtrise recherche - Polytechnique

Naga Karthik Enamundram

Doctorat - Polytechnique

Superviseur⋅e principal⋅e :

Julien Cohen-Adad

emvnagakarthik@gmail.com

Prashant Govindarajan

Doctorat - Polytechnique

Simon Guiroy

Doctorat - UdeM

Superviseur⋅e principal⋅e :

Doctorat - UdeM

David Heurtel--Depeiges

Doctorat - Polytechnique

Amir Ardalan Kalantari Dehaghi

Jerry Huang

Doctorat - UdeM

Collaborateur·rice alumni

Lola Le Breton

Doctorat - Polytechnique

Aidan Li

Maîtrise recherche - UdeM

Co-superviseur⋅e :

Postdoctorat - UdeM

Doctorat - Polytechnique

Roshan Munirathinam Sankaran Balaji

Mohamed Amine Merzouk

Postdoctorat - Polytechnique

Superviseur⋅e principal⋅e :

Stagiaire de recherche - Polytechnique

Rayen Nacef

Stagiaire de recherche - Polytechnique

Hadi NekoeiQachkanloo

Doctorat - UdeM

Nilaksh Nilaksh

Doctorat - Polytechnique

Doctorat - UdeM

Doctorat - UdeM

Postdoctorat

Visiteur de recherche indépendant

Mohammad R. Samsami

Maîtrise recherche - UdeM

Doctorat - Polytechnique

shaipraneshgci@gmail.com

Nour Shaheen

Maîtrise recherche - Polytechnique

Arjun Vaithilingam Sudhakar

Megh Thakkar

Maîtrise recherche - UdeM

Doctorat - Polytechnique

Shawn Whitfield

Collaborateur·rice de recherche

Kowen Woo

Stagiaire de recherche - Polytechnique

Abdelrahman Zayed

Doctorat - Polytechnique

Xutong Zhao

Doctorat - Polytechnique

Artem Zholus

Doctorat - Polytechnique

NeoBERT: une nouvelle frontière pour les modèles de langage encodeurs open-source

Billets de blogue

A digital picture of Bert from Sesame street, wering black trench coat and sunglasses

3 mars 2025

par

Lola Le Breton

Quentin Fournier

Sarath Chandar

Lire l'article

1 octobre 2024

Comment expliquer l’IA et s’assurer que cette explication est vraie? Les modèles mesurables de fidélité vous indiquent comment y parvenir

par

Andrea Madsen

Siva Reddy

Sarath Chandar

Lire l'article

Publications

Parity Requires Unified Input Dependence and Negative Eigenvalues in SSMs

Behnoush Khavari

Jayesh Khullar

Franccois Rivest

Recent work has shown that LRNN models such as S4D, Mamba, and DeltaNet lack state-tracking capability due to either time-invariant transiti… (voir plus)on matrices or restricted eigenvalue ranges. To address this, input-dependent transition matrices, particularly those that are complex or non-triangular, have been proposed to enhance SSM performance on such tasks. While existing theorems demonstrate that both input-independent and non-negative SSMs are incapable of solving simple state-tracking tasks, such as parity, regardless of depth, they do not explore whether combining these two types in a multilayer SSM could help. We investigate this question for efficient SSMs with diagonal transition matrices and show that such combinations still fail to solve parity. This implies that a recurrence layer must both be input-dependent and include negative eigenvalues. Our experiments support this conclusion by analyzing an SSM model that combines S4D and Mamba layers.

2025-08-10

ArXiv (prépublication)

Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models

Istabrak Abbes

Gopeshh Subbaraj

Matthew D Riemer

Nizar Islah

Benjamin Therien

Tsuguchika Tabaru

Hiroaki Kingetsu

Irina Rish

Training large language models (LLMs) typically involves pre-training on massive corpora, only to restart the process entirely when new data… (voir plus) becomes available. A more efficient and resource-conserving approach would be continual pre-training, where models are updated with new data rather than retraining from scratch. However, the introduction of new data often causes distribution shifts, leading to performance degradation on previously learned tasks. In this paper, we take a deeper look at two popular proposals for addressing this distribution shift within the continual learning literature: experience replay and gradient alignment. We consider continual pre-training of models within the Llama family of architectures at a large scale across languages with 100 billion tokens of training data in each language, finding that both replay and gradient alignment lead to more stable learning without forgetting. This conclusion holds both as we vary the model scale and as we vary the number and diversity of tasks. Moreover, we are the first to demonstrate the effectiveness of gradient alignment techniques in the context of LLM pre-training and propose an efficient implementation of meta-experience replay (MER) that imbues experience replay with the benefits of gradient alignment despite negligible compute and memory overhead. Our scaling analysis across model sizes and replay rates indicates that small rates of replaying old examples are definitely a more valuable use of compute than investing in model size, but that it is more compute efficient to scale the size of the model than invest in high rates of replaying old examples.

2025-08-03

ArXiv (prépublication)

Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models

Istabrak Abbes

Gopeshh Subbaraj

Matthew D Riemer

Nizar Islah

Benjamin Therien

Tsuguchika Tabaru

Hiroaki Kingetsu

Irina Rish

2025-08-03

ArXiv (prépublication)

Boosting LLM Reasoning via Spontaneous Self-Correction

Xutong Zhao

Tengyu Xu

Xuewei Wang

Zhengxing Chen

Di Jin

Liang Tan

Yen-Ting Lin

Zishun Yu

Zhuokai Zhao

Si-Yuan Wang

Yun He

Sinong Wang

Han Fang

MetaAI

Chen Zhu

Mila - Québec

AI Institute

Polytechnique Montréal

While large language models (LLMs) have demonstrated remarkable success on a broad range of tasks, math reasoning remains a challenging one.… (voir plus) One of the approaches for improving math reasoning is self-correction, which designs self-improving loops to let the model correct its own mistakes. However, existing self-correction approaches treat corrections as standalone post-generation refinements, relying on extra prompt and system designs to elicit self-corrections, instead of performing real-time, spontaneous self-corrections in a single pass. To address this, we propose SPOC, a spontaneous self-correction approach that enables LLMs to generate interleaved solutions and verifications in a single inference pass, with generation dynamically terminated based on verification outcomes, thereby effectively scaling inference time compute. SPOC considers a multi-agent perspective by assigning dual roles -- solution proposer and verifier -- to the same model. We adopt a simple yet effective approach to generate synthetic data for fine-tuning, enabling the model to develop capabilities for self-verification and multi-agent collaboration. We further improve its solution proposal and verification accuracy through online reinforcement learning. Experiments on mathematical reasoning benchmarks show that SPOC significantly improves performance. Notably, SPOC boosts the accuracy of Llama-3.1-8B and 70B Instruct models, achieving gains of 8.8% and 11.6% on MATH500, 10.0% and 20.0% on AMC23, and 3.3% and 6.7% on AIME24, respectively.

2025-07-07

colmweb.org/COLM/2025/Conference (accepté)

Steering Large Language Model Activations in Sparse Spaces

Reza Bayat

Ali Rahimi-Kalahroudi

Mohammad Pezeshki

Pascal Vincent

2025-07-07

colmweb.org/COLM/2025/Conference (accepté)

CADmium: Fine-Tuning Code Language Models for Text-Driven Sequential CAD Design

Prashant Govindarajan

Davide Baldelli

Jay Pathak

Quentin Fournier

Computer-aided design (CAD) is the digital construction of 2D and 3D objects, and is central to a wide range of engineering and manufacturin… (voir plus)g applications like automobile and aviation. Despite its importance, CAD modeling remains largely a time-intensive, manual task. Recent works have attempted to automate this process with small transformer-based models and handcrafted CAD sequence representations. However, there has been little effort to leverage the potential of large language models (LLMs) for sequential CAD design. In this work, we introduce a new large-scale dataset of more than 170k CAD models annotated with high-quality, human-like descriptions generated with our pipeline based on GPT-4.1. Using this dataset, we fine-tune powerful code-LLMs to generate CAD sequences represented in a JSON-based format from natural language descriptions, demonstrating the viability and effectiveness of this approach for text-conditioned CAD generation. Because simple metrics often fail to reflect the quality of generated objects, we introduce geometric and topological metrics based on sphericity, mean curvature, and Euler characteristic to provide richer structural insights. Our experiments and ablation studies on both synthetic and human-annotated data demonstrate that CADmium is able to automate CAD design, drastically speeding up the design of new objects. The dataset, code, and fine-tuned models are available online.

2025-07-01

arXiv (publié)

Optimizers Qualitatively Alter Solutions And We Should Leverage This

Razvan Pascanu

Clare Lyle

Ionut-Vlad Modoranu

Naima Elosegui Borras

Dan Alistarh

Petar Veličković

Soham De

James Martens

Due to the nonlinear nature of Deep Neural Networks (DNNs), one can not guarantee convergence to a unique global minimum of the loss when us… (voir plus)ing optimizers relying only on local information, such as SGD. Indeed, this was a primary source of skepticism regarding the feasibility of DNNs in the early days of the field. The past decades of progress in deep learning have revealed this skepticism to be misplaced, and a large body of empirical evidence shows that sufficiently large DNNs following standard training protocols exhibit well-behaved optimization dynamics that converge to performant solutions. This success has biased the community to use convex optimization as a mental model for learning, leading to a focus on training efficiency, either in terms of required iteration, FLOPs or wall-clock time, when improving optimizers. We argue that, while this perspective has proven extremely fruitful, another perspective specific to DNNs has received considerably less attention: the optimizer not only influences the rate of convergence, but also the qualitative properties of the learned solutions. Restated, the optimizer can and will encode inductive biases and change the effective expressivity of a given class of models. Furthermore, we believe the optimizer can be an effective way of encoding desiderata in the learning process. We contend that the community should aim at understanding the biases of already existing methods, as well as aim to build new optimizers with the explicit intent of inducing certain properties of the solution, rather than solely judging them based on their convergence rates. We hope our arguments will inspire research to improve our understanding of how the learning process can impact the type of solution we converge to, and lead to a greater recognition of optimizers design as a critical lever that complements the roles of architecture and data in shaping model outcomes.

2025-07-01

arXiv (publié)

Small Encoders Can Rival Large Decoders in Detecting Groundedness

Istabrak Abbes

Gabriele Prato

Quentin Fournier

Fernando Rodriguez

Alaa Boukhary

Adam Elwood

Augmenting large language models (LLMs) with external context significantly improves their performance in natural language processing (NLP) … (voir plus)tasks. However, LLMs struggle to answer queries reliably when the provided context lacks information, often resorting to ungrounded speculation or internal knowledge. Groundedness - generating responses strictly supported by the context - is essential for ensuring factual consistency and trustworthiness. This study focuses on detecting whether a given query is grounded in a document provided in context before the costly answer generation by LLMs. Such a detection mechanism can significantly reduce both inference time and resource consumption. We show that lightweight, task specific encoder models such as RoBERTa and NomicBERT, fine-tuned on curated datasets, can achieve accuracy comparable to state-of-the-art LLMs, such as Llama3 8B and GPT4o, in groundedness detection while reducing inference latency by orders of magnitude. The code is available at : https://github.com/chandarlab/Hallucinate-less

2025-06-26

ArXiv (prépublication)

Revisiting Laplacian Representations for Value Function Approximation in Deep RL

Rishav

Yash Chandak

Samira Ebrahimi Kahou

Doina Precup

Proto-value functions (PVFs) introduced Laplacian embeddings as an effective feature basis for value-function approximation; however, their … (voir plus)utility remained limited to small, fully known state spaces. Recent work has scaled Laplacian embeddings to high-dimensional inputs, using them for reward shaping and option discovery in goal-directed tasks, yet only as auxiliary signals, rather than directly using them as features for value functions. In this paper, we learn Laplacian eigenvectors online and employ them as features for Q-learning in 23 Atari games. We empirically demonstrate that these online–learned embeddings substantially improve model-free RL in large, high-dimensional domains. We demonstrate that enriching state representations with action embeddings yields additional gains under both behavior-policy and uniform-random policies. Additionally, we introduce the Fusion architecture, which augments the representation with useful inductive bias at the embedding level. To assess the usefulness of each embedding used in the Fusion architecture, we use Shapley values analysis.

2025-06-22

rl-conference.cc/RLC/2025/Workshop/IBRL (publié)

Monitoring morphometric drift in lifelong learning segmentation of the spinal cord

Enamundram Naga Karthik

Sandrine Bédard

Jan Valosek

Julien Cohen-Adad

Christoph S. Aigner

Elise Bannier

Josef Bednařík

Virginie Callot

Anna Combes

Armin Curt

Gergely David

Falk Eippert

Lynn Farner

Michael G Fehlings

Patrick Freund

Tobias Granberg

Cristina Granziera

Rhscir Network Imaging Group

Ulrike Horn

Tomáš Horák … (voir 36 de plus)

Suzanne Humphreys

Markus Hupp

Anne Kerbrat

Nawal Kinany

Shannon Kolind

Anna Lebret

Petr Kudlička

Lisa Eunyoung Lee

Caterina Mainero

Allan R. Martin

Megan McGrath

Govind Nair

Kristin P. O’Grady

Jiwon Oh

Russell Ouellette

Nikolai Pfender

Dario Pfyffer

P. Pradat

Alexandre Prat

Emanuele Pravatà

Daniel S. Reich

Ilaria Ricchi

Naama Rotem-Kohavi

Simon Schading-Sassenhausen

Maryam Seif

Andrew C. Smith

Seth Aaron Smith

Grace Sweeney

Roger Tam

Anthony Traboulsee

Constantina A. Treaba

Charidimos Tsagkas

Zachary Vavasour

Dimitri Van De Ville

Kenneth A. Weber

2025-06-12

lifelong-ml.cc/CoLLAs/2025/Workshop_Track (publié)

NovoMolGen: Rethinking Molecular Language Model Pretraining

Kamran Chitsaz

Roshan Balaji

Quentin Fournier

Nirav Pravinbhai Bhatt

2025-06-11

ICML.cc/2025/Workshop/GenBio (poster)

V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning

Mahmoud Assran

Adrien Bardes

David Fan

Quentin Garrido

Russell Howes

Mojtaba Komeili

Matthew J. Muckley

Ammar Rizvi

Claire Roberts

Koustuv Sinha

Artem Zholus

Sergio Arnaud

Abha Gejji

Ada Martin

Francois Robert Hogan

Daniel Dugas

Piotr Bojanowski

Vasil Khalidov

Patrick Labatut

Francisco Massa … (voir 13 de plus)

Marc Szafraniec

K. Krishnakumar

Ying Li

Xiaodong Ma

Franziska Meier

Yann LeCun

Michael Rabbat

Nicolas Ballas

Fair at Meta

Mila - Québec

AI Institute

Polytechnique Montréal

A major challenge for modern AI is to learn to understand the world and learn to act largely by observation. This paper explores a self-supe… (voir plus)rvised approach that combines internet-scale video data with a small amount of interaction data (robot trajectories), to develop models capable of understanding, predicting, and planning in the physical world. We first pre-train an action-free joint-embedding-predictive architecture, V-JEPA 2, on a video and image dataset comprising over 1 million hours of internet video. V-JEPA 2 achieves strong performance on motion understanding (77.3 top-1 accuracy on Something-Something v2) and state-of-the-art performance on human action anticipation (39.7 recall-at-5 on Epic-Kitchens-100) surpassing previous task-specific models. Additionally, after aligning V-JEPA 2 with a large language model, we demonstrate state-of-the-art performance on multiple video question-answering tasks at the 8 billion parameter scale (e.g., 84.0 on PerceptionTest, 76.9 on TempCompass). Finally, we show how self-supervised learning can be applied to robotic planning tasks by post-training a latent action-conditioned world model, V-JEPA 2-AC, using less than 62 hours of unlabeled robot videos from the Droid dataset. We deploy V-JEPA 2-AC zero-shot on Franka arms in two different labs and enable picking and placing of objects using planning with image goals. Notably, this is achieved without collecting any data from the robots in these environments, and without any task-specific training or reward. This work demonstrates how self-supervised learning from web-scale data and a small amount of robot interaction data can yield a world model capable of planning in the physical world.

2025-06-11

ArXiv (prépublication)