Sarath Chandar

istabrak.abbes@mila.quebec

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

Istabrak Abbes

Doctorat - UdeM

Alex Aselstyne

Maîtrise recherche - Polytechnique

alexander.aselstyne@mila.quebec

davide.baldelli@mila.quebec

Davide Baldelli

Doctorat - Polytechnique

Co-superviseur⋅e :

Amal Zouaq

diego.cerda-mardini@mila.quebec

Milan Bhan

Collaborateur·rice de recherche

milan.bhan@mila.quebec

Diego Cerda Mardini

Maîtrise recherche - McGill

Naga Karthik Enamundram

Doctorat - Polytechnique

Superviseur⋅e principal⋅e :

Julien Cohen-Adad

emvnagakarthik@gmail.com

prashant.govindarajan@mila.quebec

Prashant Govindarajan

Doctorat - Polytechnique

maryam.hashemzadeh@mila.quebec

Simon Guiroy

Doctorat - UdeM

Superviseur⋅e principal⋅e :

Doctorat - UdeM

david.heurtel-depeiges@mila.quebec

David Heurtel--Depeiges

Doctorat - Polytechnique

Jerry Huang

Doctorat - UdeM

jerry.huang@mila.quebec

Kanishk Jain

Doctorat - UdeM

Co-superviseur⋅e :

Ross Goroshin

kanishk.jain@mila.quebec

Saurav Jha

Postdoctorat - Polytechnique

saurav.jha@mila.quebec

Amir Ardalan Kalantari Dehaghi

Doctorat - UdeM

Doctorat - Polytechnique

lola.lebreton@mila.quebec

ekaterina.lobacheva@mila.quebec

Aidan Li

Maîtrise recherche - UdeM

Co-superviseur⋅e :

Postdoctorat - UdeM

Doctorat - Polytechnique

pranshu.malviya@mila.quebec

Roshan Munirathinam Sankaran Balaji

Collaborateur·rice de recherche - Polytechnique

roshan.munirathinam@mila.quebec

nilaksh.nilaksh@mila.quebec

Hadi NekoeiQachkanloo

Doctorat - UdeM

nekoeihe@mila.quebec

Nilaksh Nilaksh

Doctorat - Polytechnique

Doctorat - UdeM

darshan.patil@mila.quebec

Linda Peinthiere

Collaborateur·rice de recherche - Polytechnique Montreal

uphro_1@hotmail.com

mathieu.reymond@mila.quebec

Yann Pernot

Maîtrise recherche - Polytechnique

yann.pernot@mila.quebec

Mathieu Reymond

Collaborateur·rice alumni

Shaipranesh Senthilkumar

Doctorat - Polytechnique

shaipraneshgci@gmail.com

Arjun Vaithilingam Sudhakar

Nour Shaheen

Maîtrise recherche - Polytechnique

Superviseur⋅e principal⋅e :

Amine Mhedhbi

nour.shaheen@mila.quebec

Doctorat - Polytechnique

achille.sowa@mila.quebec

Anabel Tan

Postdoctorat - UdeM

anabel.tan@mila.quebec

Megh Thakkar

Maîtrise recherche - UdeM

megh.thakkar@mila.quebec

Doctorat - Polytechnique

arjun.vaithilingam-sudhakar@mila.quebec

Abdelrahman Zayed

Doctorat - Polytechnique

zayedabd@mila.quebec

Xutong Zhao

Doctorat - Polytechnique

xutong.zhao@mila.quebec

Artem Zholus

Doctorat - Polytechnique

artem.zholus@mila.quebec

Optimiser la conception CAO grâce aux LLM

Billets de blogue

19 décembre 2025

par

Prashant Govindarajan

Davide Baldelli

Quentin Fournier

Sarath Chandar

Lire l'article

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

3 mars 2025

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

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

Addressing Sparse-Rewards in RL with Scalable Hierarchical Novel Eigen Options

Isabeau Prémont-Schwarz

Temporally extended exploration via graph Laplacian-based options is a promising approach to sparse-reward reinforcement learning (RL), but … (voir plus)existing methods either do not explicitly target novelty or fail to scale to pixel-based domains under function approximation. Novel Exploration via Orthogonality (NEO) addresses the first issue by constructing options that navigate from highly visited regions toward less visited ones, yet prior results were limited to settings where exact eigenvectors can be computed. We present a scalable extension of NEO to pixel-based domains, built on three contributions. First, we use a novelty-weighted continuous Laplacian graph-drawing objective, which enables RL with continuous observations. Second, we embed the resulting eigen-potential options within a hierarchical reinforcement learning framework, enabling coherent temporally extended behavior. Third, we observe that learned eigen-potential rewards are directional but locally unreliable under online approximation; we therefore augment each option reward with a novelty bonus, a novel design idea that proves essential for stabilizing option learning while preserving novelty-directed exploration. Together, these contributions yield stronger and more persistent exploration, enabling longer option rollouts and better access to hard-to-reach novel states. Empirically, our method significantly outperforms both the prior scalable Laplacian-option baseline and a direct extension of NEO on sparse-reward benchmarks under a fixed budget. On Montezuma's Revenge, our best variant achieves approximately 1.8x higher return than both baselines. On Venture, both baselines yield returns near zero, whereas our method achieves a return of 1135. Across seven hard ProcGen games, our method achieves approximately 3.5x and 5.6x higher aggregate normalized return than the two baselines, respectively.

2026-07-20

RLBRew @ Reinforcement Learning Conference (publié)

A systematic analysis of machine learning pipelines for robust antimicrobial resistance prediction

Alex Aselstyne

Enamundram Naga Karthik

Meriem El Azami

Romain Pogorelcnik

Quentin Fournier

Abstract Motivation Antimicrobial resistance (AMR) has been identified as a top global public health threat. Accurate AMR phenotype predicti… (voir plus)on from whole-genome sequencing data is an essential tool for accelerating clinical decision-making and mitigating resistance spread. Although many previous works have explored the use of tree-based machine learning (ML) models to predict resistance, the field lacks a systematic evaluation of the training pipeline across a variety of pathogenic species and antibiotics. Results Using nine clinically relevant species–antibiotic combinations from the NCBI antimicrobial susceptibility testing database, we present a detailed analysis of the ML pipeline and identify key factors affecting model performance and evaluation. We begin by relabelling all isolates using current CLSI minimum inhibitory concentration breakpoints to resolve inconsistencies and increase available data, resulting in up to a 19% label swap and 56% data enlargement per species– antibiotic combination. We identify several key training parameters including k -mer length, which can increase classification F1 scores by over 20 points compared to commonly used k -values, feature matrix truncation, which can induce polynomial time reductions with limited performance reduction, and ML model class. By comparing 5-fold cross-validation with evaluation on an unseen clinical dataset, we show that random cross-validation splits—often criticized as overly optimistic—can act as a strong proxy for downstream clinical performance, yielding closer F1 scores than phylogeny-aware splits in all cases. We finally present an interpretability study which shows that over 95% of k -mers used by our models are associated with identifiable genomic features. Our results highlight the importance of feature design, evaluation protocol, and biological analysis in genomic AMR prediction, and support tree-based models as a robust and interpretable method. Availability and implementation Python code is made freely available: https://github.com/chandar-lab/amr-pred

2026-07-07

bioRxiv (prépublication)

Loss Smoothing for Stable Adaptation Under Distribution Shift

In settings such as fine-tuning and reinforcement learning, neural networks are often adapted under distribution shift. Standard adaptation … (voir plus)methods typically optimize the target objective directly, inducing an abrupt change from the source training objective. This abrupt transition can distort learned representations, including features that may still be useful for the new task. We investigate whether a more gradual transition can improve adaptation. We propose loss smoothing, a simple approach that interpolates between the source and target training objectives at the start of adaptation. This smooth transition helps to preserve useful features from the source distribution while still enabling the model to specialize to the target distribution. Across controlled supervised shifts, pretrained vision adaptation, offline-to-online and online reinforcement learning, and language model fine-tuning, we find that loss smoothing consistently improves performance, suggesting that smoother objective transitions are a broadly useful tool for model adaptation.

2026-06-30

arXiv (prépublication)

NeuroFaith: Evaluating Mechanistic Faithfulness of LLM Free Text Self-Explanation at the Concept Level

Milan Bhan

Jean-Noël Vittaut

Nicolas Chesneau

Marie-Jeanne Lesot

Large Language Models (LLMs) can generate plausible free text self-explanations to justify their answers. However, these natural language ex… (voir plus)planations may not accurately reflect the model's actual reasoning process, indicating a lack of faithfulness. Existing faithfulness evaluation methods rely primarily on behavioral tests or computational block analysis without examining the semantic content of internal neural representations. This paper proposes NeuroFaith, a flexible framework that measures the faithfulness of LLM free text self-explanation by identifying key concepts within explanations and mechanistically testing whether these concepts actually influence the model's predictions. We show the versatility of NeuroFaith across 2-hop reasoning and classification tasks. Additionally, we develop a linear faithfulness probe based on NeuroFaith to detect unfaithful self-explanations from representation space and improve faithfulness through steering. NeuroFaith provides a principled approach to evaluating and enhancing the faithfulness of LLM free text self-explanations, addressing critical needs for trustworthy AI systems.

2026-06-10

ICML.cc/2026/Workshop/Mech_Interp (poster)

Failed Reasoning Traces Tell You What Is Fixable (But Not by Reading Them)

When post-trained language models fail on reasoning problems, the common test-time-scaling response is to spend more compute on additional a… (voir plus)ttempts, and the failed traces play no further role. We argue this discards a crucial signal; some failures come from unlucky sampling, where more rollouts help, while others are structural and resist resampling regardless of budget. We propose that failed traces encode recoverability structure: the inference-time signature of which test-time interventions can rescue a given failure. Three problem-level trajectory features, derived from the structure of available interventions, recover this structure from the distributional signature of failed rollouts, not their text. They cluster failures into stable regimes, characterize the failure topography of different post-training methods (

2026-06-02

arXiv (prépublication)

CrysTune: Crystal Generation via Fine-Tuning of Large Language Models on Wyckoff Representations

Prashant Govindarajan

KISHALAY DAS

The discovery of novel materials is essential for driving scientific and technological breakthroughs. Recent work has explored fine-tuning l… (voir plus)arge language models (LLMs) for autoregressive crystal generation, but the ideal representation and training strategies for symmetry-based inductive biases remain unclear. We propose CrysTune, a class of LLMs fine-tuned on Wyckoff representations of crystals with two auxiliary tasks: canonicalization and template prediction. CrysTune shows competitive performance and improved stability-related metrics relative to LLMs trained on standard string-encoded representations. We further use these models as initial policies for reinforcement learning (RL) fine-tuning to optimize stability, validity, uniqueness, novelty, and diversity. RL-trained policies produce more valid and metastable crystals, while introducing novelty and diversity trade-offs. We also explore crystal system conditioning, showing that RL-trained policies produce a higher proportion of crystals matching the target condition.

2026-05-29

AI4Science @ International Conference on Machine Learning (poster)

Probabilistic Calibration Is a Trainable Capability in Language Models

Sruthi Kuriakose

Language models are increasingly used in settings where outputs must satisfy user-specified randomness constraints, yet their generation pro… (voir plus)babilities are often poorly calibrated to those targets. We study whether this capability can be improved directly through fine-tuning. Concretely, we fine-tune language models on synthetic prompts that require sampling from mathematical distributions, and compare two Calibration Fine-Tuning variants: a soft-target method that converts the desired output distribution into trie-derived next-token targets, and a hard-target method that trains on sampled completions from the same target distribution. Across 12 models spanning four families, both methods substantially improve structured-sampling fidelity on held-out distribution families and unseen parameter settings, showing that probabilistic calibration is a trainable capability. Under our selected training configurations, the two methods exhibit different empirical profiles: hard-target fine-tuning is often strongest on structured numeric sampling, while soft-target fine-tuning performs better on broader stochastic generation benchmarks, including open-ended random generation, multiple-choice answer-position balancing, and NoveltyBench. The gains sometimes reduce downstream capability, especially arithmetic reasoning, with costs varying by model. Overall, our results show that probabilistic calibration can be improved through fine-tuning, with our hard-target configuration favoring exact numeric fidelity and our soft-target configuration favoring broader stochastic transfer. Code is available at https://github.com/chandar-lab/calibration-finetuning.

2026-05-11

arXiv (prépublication)

TAPNext++: What's Next for Tracking Any Point (TAP)?

Sebastian Jung

Artem Zholus

Martin Sundermeyer

Carl Doersch

Ross Goroshin

David Joseph Tan

Rudolph Triebel

Federico Tombari

Tracking-Any-Point (TAP) models aim to track any point through a video which is a crucial task in AR/XR and robotics applications. The recen… (voir plus)tly introduced TAPNext approach proposes an end-to-end, recurrent transformer architecture to track points frame-by-frame in a purely online fashion -- demonstrating competitive performance at minimal latency. However, we show that TAPNext struggles with longer video sequences and also frequently fails to re-detect query points that reappear after being occluded or leaving the frame. In this work, we present TAPNext++, a model that tracks points in sequences that are orders of magnitude longer while preserving the low memory and compute footprint of the architecture. We train the recurrent video transformer using several data-driven solutions, including training on long 1024-frame sequences enabled by sequence parallelism techniques. We highlight that re-detection performance is a blind spot in the current literature and introduce a new metric, Re-Detection Average Jaccard (

2026-04-11

arXiv (prépublication)

Emergent Reasoning via Recursive Latent Reinforcement Pretraining

Large language models (LLMs) often rely on explicit chain-of-thought (CoT) traces to solve multi-step reasoning problems, but these traces i… (voir plus)ncrease inference cost, expose brittle prompt dependence, and complicate training objectives. We study an alternative: \emph{latent deliberation} implemented as a small recurrent refinement module that performs multiple internal ``thinking`` steps while keeping the external sequence length fixed. We introduce \textbf{Recursive Latent Reinforcement Pretraining (RLRP)}, a training recipe that augments a base causal LLM with a shared latent head executed for

2026-03-04

LLM_Reasoning @ International Conference on Learning Representations (publié)

Is Depth Heterogeneity a Barrier to Model Merging?

Model merging offers a way to combine the capabilities of several networks at test time without retraining or additional finetuning, but mos… (voir plus)t merging methods assume identical architectures. Depth differences are commonly viewed as a major obstacle because they remove clear layer correspondences. We test this assumption by merging residual networks that differ only in depth, using a simple training-free pipeline based on identity expansion and permutation alignment. Across both same-task and multitask image classification experiments, heterogeneous merges closely match homogeneous ones. The results suggest that, for residual networks, depth mismatch is not the main barrier to effective model merging, and that the main difficulty in model merging comes from aligning independently trained weights in a homogeneous setting.

2026-02-28

TTU_Main_Track @ International Conference on Learning Representations (publié)

Loss Smoothing for Continual Adaptation

Neural networks are often adapted in nonstationary data distributions settings where the objective is to optimize performance on the current… (voir plus) task, and preserving accuracy on previous tasks is not required. As a result, existing methods primarily focus on improving plasticity, while stability is largely studied in the context of continual learning. In this work, we examine whether preserving stability can also be beneficial in model adaptation settings where past-task performance is irrelevant. We propose a simple loss smoothing approach that encourages selective adaptation by preserving task-shared features while modifying task-inconsistent ones. We evaluate our method on continual supervised model adaptation benchmarks and reinforcement learning benchmarks, and show that promoting representational stability during adaptation can improve performance across settings.

2026-02-27

CAO @ International Conference on Learning Representations (poster)

CoPeP: Benchmarking Continual Pretraining for Protein Language Models

Protein language models (pLMs) have recently gained significant attention for their ability to uncover relationships between sequence, struc… (voir plus)ture, and function from evolutionary statistics, thereby accelerating therapeutic drug discovery. These models learn from large protein databases that are continuously updated by the biology community and whose dynamic nature motivates the application of continual learning, not only to keep up with the ever-growing data, but also as an opportunity to take advantage of the temporal meta-information that is created during this process. As a result, we introduce the Continual Pretraining of Protein Language Models (CoPeP) benchmark, a novel benchmark for evaluating continual learning approaches on pLMs. Specifically, we curate a sequence of protein datasets derived from the UniProt Knowledgebase spanning a decade and define metrics to assess pLM performance across 31 protein understanding tasks. We evaluate several methods from the continual learning literature, including replay, unlearning, and plasticity-based methods, some of which have never been applied to models and data of this scale. Our findings reveal that incorporating temporal meta-information improves perplexity by up to 7% even when compared to training on data from all tasks jointly. Moreover, even at scale, several continual learning methods outperform naive continual pretraining. The CoPeP benchmark offers an exciting opportunity to study these methods at scale in an impactful real-world application.

2026-02-26

arXiv (prépublication)