Publications

Improving column complementarity in a restricted master heuristic with a GRASP-guided completion: Application to the vehicle routing problem with stochastic demands
Gaël Reynal
Quentin Cappart
Guy Desaulniers
Louis-Martin Rousseau
2026-06-04
Computers & Operations Research (publié)
doi.org
ULMShare: A Large-Scale In Vivo Ultrasound Localization Microscopy Dataset for Microvascular Imaging
Brice Rauby
Nin Ghigo
Gerardo Ramos-Palacios
Alexis Leconte
Stephen A. Lee
Alice Wu
Paul Xing
Oleksandra Gulenko
Louis Caron
Antoine Malescot
Eric Martineau
Jonathan Porée
Maxime Gasse
Ravi L. Rungta
Abbas F. Sadikot
Jean Provost
Ultrasound Localization Microscopy (ULM) enables microscopic imaging of the cerebral microvasculature in vivo, but relies on a multi-stage p… (voir plus)rocessing pipeline in which acquisition settings and reconstruction processes strongly influence the final output. Existing public datasets remain sparse, restricting rigorous evaluation and slowing progress in algorithm development, including emerging machine-learning approaches, which by design require large quantities of data to be robust and reliable. We introduce \textbf{ULMShare}, an open-access dataset of 99 whole-brain transcranial ULM acquisitions from 61 healthy mice (36 females, 22 males, 3 unknown; mean age:
2026-06-04
arXiv (prépublication)
doi.org
arxiv.org
Unsupervised Continual Clustering via Forward-Backward Knowledge Distillation
Mohammadreza Sadeghi
Sareh Soleimani
Zihan Wang
Narges Armanfard
Unsupervised Continual Learning (UCL) aims to enable neural networks to learn sequential tasks without labels or access to past data. A majo… (voir plus)r challenge in this setting is Catastrophic Forgetting, where models forget previously learned tasks upon learning new ones. This challenge is amplified in UCL due to the absence of labels to guide learning and memory retention. Existing mitigation strategies, such as knowledge distillation and replay buffers, often raise memory and privacy concerns. Moreover, current UCL methods largely overlook clustering-specific objectives. To fill this gap, we introduce Unsupervised Continual Clustering (UCC) and propose Forward-Backward Knowledge Distillation for Continual Clustering (FBCC). FBCC employs a continual teacher network with a clustering projector and lightweight task-specific students. Through a dual-phase forward-backward distillation process, the teacher learns new clusters while preserving previously discovered cluster structure without storing past data. FBCC represents a pioneering approach to UCC, demonstrating improved clustering performance across sequential tasks. Experiments on four benchmark datasets demonstrate that FBCC consistently outperforms existing continual learning baselines in clustering accuracy while significantly reducing catastrophic forgetting.
2026-06-04
arXiv (prépublication)
doi.org
arxiv.org
Video-Based Prediction of In-Flight Particle Characteristics in Atmospheric Plasma Spraying
Abhijeet Praveen
Sareh Soleimani
Cormac Cureton
Aman Sidhu
Kintak Raymond Yu
Cristian Cojocaru
Narges Armanfard
Atmospheric plasma spraying (APS) is a widely used coating process in which in-flight particle temperature and velocity strongly influence c… (voir plus)oating quality. However, these particle characteristics are difficult to monitor continuously during operation, motivating the development of non-invasive data-driven diagnostic methods. In this work, we investigate the predictive potential of high-speed video observations of the plasma plume for estimating in-flight particle characteristics in APS. We introduce three different video-derived feature representations and evaluate them using Tabular Prior-Data Fitted Networks (TabPFN), convolutional neural networks (CNN), and classical regression baselines including Random Forest, Gradient Boosting, Support Vector Regression, and XGBoost. Experiments are conducted using grouped leave-one-out cross-validation on 126 labeled pre- and post-spray video recordings from 63 APS spray runs. Across the engineered feature experiments, TabPFN achieves the most consistent performance for temperature prediction, reaching R2 = 0.86 using the combined feature representation. CNN models particularly perform stronger for velocity prediction, achieving R2 of 0.81. In addition, we evaluate models operating directly on raw video frames using pretrained CNNs and find that the highest performance is achieved by a pretrained CNN with a regression head with R2 of 0.90 and 0.82 for temperature and velocity, respectively. The results demonstrate that video-derived plume information provides a promising and scalable foundation for non-invasive APS diagnostics and real-time process monitoring.
2026-06-04
arXiv (prépublication)
doi.org
arxiv.org
Human Adults and LLMs as Scientists: Who Benefits from Active Exploration?
Mandana Samiei
Eunice Yiu
Anthony GX-Chen
Dongyan Lin
Jocelyn Shen
Blake A. Richards
Alison Gopnik
Doina Precup
A long-standing finding in the causal learning literature is that adults struggle to identify conjunctive causal rules, where an effect requ… (voir plus)ires the simultaneous presence of multiple causes, while performing better in disjunctive settings. However, most demonstrations of this ``conjunctive handicap'' rely on passive observation paradigms with limited evidence, where learners have no control over evidence generation. This paper asks whether this bias persists when adults are granted agency through active exploration. Using a modified ``blicket detector'' task, adult participants freely intervened to identify causal objects under conjunctive or disjunctive rule structures. We show that active exploration substantially improves adults' conjunctive causal reasoning, although conjunctive rules still require more tests to infer than disjunctive rules. We further compare human performance to a range of large language models in the same setting. While some state-of-the-art models approach human-level performance on hypothesis inference accuracy, they often exhibit less efficient exploration strategies and similar conjunctive-disjunctive performance gaps.
2026-06-03
arXiv (prépublication)
doi.org
arxiv.org
Noticing the Watcher: LLM Agents Can Infer CoT Monitoring from Blocking Feedback
Thomas Jiralerspong
Flemming Kondrup
Yoshua Bengio
Chain-of-thought (CoT) monitoring provides oversight into model reasoning, but its effectiveness assumes models do not know they are being w… (voir plus)atched. We ask whether reasoning agents can autonomously infer that their supposedly private chain of thought is under surveillance, and whether this awareness leads to strategic evasion, without any explicit training or instructions to do so. In a multi-episode agentic framework, models pursue both a primary task and a concealed side task while being told their reasoning is private; a hidden CoT monitor blocks episodes when suspicious reasoning is detected. We find that frontier models can deduce the existence of this monitor purely from blocking feedback, with the most capable models reaching confident belief that their thinking is observed in up to 19\% of episodes. This awareness scales with model capability and, in rare cases, escalates to explicit intent to suppress reasoning about the side task. However, models that form this intent uniformly fail to execute it, openly reasoning about their concealed objectives in the very next episode. This intent–capability gap is reassuring for current deployment, but the autonomous emergence of both monitoring awareness and evasion intent suggests that CoT monitoring is not a permanently reliable safeguard.
2026-06-03
PhilML @ International Conference on Machine Learning (poster)
doi.org
openreview.net
Representation Learning Enables Scalable Multitask Deep Reinforcement Learning
Johan Obando-Ceron
Lu Li
Scott Fujimoto
Pierre-Luc Bacon
Aaron Courville
Pablo Samuel Castro
Scaling reinforcement learning (RL) to diverse multitask settings remains a central challenge. While recent advances in model-based RL achie… (voir plus)ve strong performance, they rely on planning and complex training pipelines, making it unclear which components are essential for scalability. We revisit this question and argue that the primary driver of scalable multitask RL is not model-based control, but \emph{representation learning}. In particular, we show that combining predictive, model-based representations with high-capacity value function approximation is sufficient to achieve strong performance, even without planning. We evaluate a simple model-free algorithm, MR.Q, coupled with auxiliary predictive objectives into a scalable actor-critic architecture. This approach outperforms a recent world-model-based method and a range of deep RL baselines across a diverse suite of multitask continuous control tasks, while significantly reducing computational overhead and improving wall-clock efficiency. We observe consistent improvements with increased model capacity and show through ablations that predictive representation learning is critical for performance.
2026-06-03
arXiv (prépublication)
doi.org
arxiv.org
SpeechJBB: Probing Safety Alignment and Comprehension in Large Audio Language Models under Code-Switched Speech
Virginia Ceccatelli
Yejin Jeon
David Ifeoluwa Adelani
Large audio language models (LALMs) are increasingly deployed in real-world applications, yet their safety alignment is still primarily eval… (voir plus)uated on monolingual, text-based harmful prompts. This leaves their generalizability under multilingual and spoken settings, particularly code-switched speech, largely underexplored. To address this gap, we introduce SpeechJBB, an audio jailbreak dataset for benchmarking across multiple state-of-the-art LALMs. The extent of safety weaknesses is further probed by introducing an augmented setting where phonologically plausible pseudo-words are inserted around safety-critical terms to simulate localized obfuscation. Across models, code-switched harmful audio yields substantially high jailbreak success rates (JSR), with non-English monolingual and non-English code-switched pairs exhibiting the highest attack success. Pseudo-word insertion further reduces refusal rates, which demonstrates that natural-sounding obfuscation can effectively bypass safety policies.
2026-06-03
arXiv (prépublication)
doi.org
arxiv.org
Brain states recur across diverse narrative contexts during longitudinal viewing
Yibei Chen
Matin Ghavami
Marie St‐Laurent
Lune Bellec
Satrajit S. Ghosh
Abstract What does the brain do during the continuous, varied experience of watching a story unfold? One account holds that the brain traver… (voir plus)ses a finite repertoire of recurring states, but whether that repertoire is a stable property of the individual or is reshaped by each new experience has not been tested across diverse naturalistic content within the same person. We characterized the dynamic brain-state repertoire in six individuals who watched the television series Friends across its six seasons during fMRI (up to ∼146 episodes, ∼54 hours per person). For each individual we fit a sticky hierarchical Dirichlet process hidden Markov model across all episodes, discovering brain states (recurring whole-brain activity patterns with characteristic coupling) without pre-specifying their number. Each individual’s brain visited roughly forty-five states arrayed along a continuous recurrence gradient, from states active in nearly every episode to episode-specific ones, with no sharp division between them. The repertoire was heterogeneous in why its states recurred: a minority locked to scan-run structure, the majority remaining eligible for content. Transitions were organized by the functional-connectivity similarity between states (per-individual Spearman ρ = 0.33–0.55) and, in most individuals, respected resting-state network boundaries. Episode content was associated with which states the brain occupied moment to moment. The recurrence ordering discovered in Friends transferred to state occupancy during other social-narrative films (five of six individuals) and attenuated as stimuli departed from that class, weakening for visual-only reading and audio-only listening. Across diverse narrative experience, the dynamic repertoire is a property of the individual: content varies which states are visited and when, not which states exist.
2026-06-02
bioRxiv (prépublication)
doi.org
Failed Reasoning Traces Tell You What Is Fixable (But Not by Reading Them)
Nizar Islah
Istabrak Abbes
Irina Rish
Sarath Chandar
Eilif B. Muller
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)
doi.org
arxiv.org
Learned Subspace Compression for Communication-Efficient Pipeline Parallelism
Paul Janson
Edouard Oyallon
Eugene Belilovsky
Pipeline parallelism enables training of large language models that exceed single-device memory, yet inter-stage activation communication be… (voir plus)comes the dominant bottleneck when trained on low-bandwidth networks. Recent work in this area has proposed using fixed orthogonal projections to compress activations. However, this still results in a significant performance degradation and requires a number of non-standard adaptations to constrain the optimization. A natural alternative is to learn a low rank projection for each pipeline stage, however maintaining the necessary orthogonality of these projectors during training remains a challenge. We present Manifold Aware Projection Learning (MAPL), a method that treats inter-stage compression as a learnable orthogonal projection under explicit Stiefel manifold (orthogonal matrices) constraints. Rather than prescribing a fixed global subspace, MAPL lets each pipeline stage discover and continuously adapt its own task-optimal compression subspace via manifold-constrained steepest descent. To recover token-specific signals at stage boundaries, we introduce per-stage factorized anchor embeddings that allow for full-rank activation reconstruction with negligible communication overhead. We further show that we can incorporate residual vector quantization after projection with a streaming codebook synchronization protocol that amortizes dictionary communication. Across LLaMA models from 150M to 1B parameters we show that MAPL can be easily applied to the existing pipeline and can achieve high compression with neglibile performance degradation with a drastically improved tradeoffs in performance vs. compression compared to Subspace Networks.
2026-06-02
arXiv (prépublication)
doi.org
arxiv.org
Learning Admissible Heuristics via Cost Partitioning
Hugo Barral
Quentin Cappart
Marie-José Huguet
Sylvie Thiébaux
Admissible heuristics are essential for optimal planning, yet learning them remains challenging due to the risk of overestimation. Cost part… (voir plus)itioning combines multiple abstraction heuristics while preserving admissibility, but computing optimal partitions online is expensive. We propose a framework that learns to infer admissible cost partitions by leveraging the Lagrangian dual equivalence between cost partitioning and multiplier prediction. Planning states and patterns are encoded as labelled graphs, and an action-centric variant of the Weisfeiler-Leman algorithm extracts structural feature vectors. A deep architecture with axial self-attention and a softmax output layer maps these features to cost weights that satisfy the partition constraints by construction, ensuring admissibility. Experiments demonstrate reduced node expansions compared to suboptimal partitioning baselines while maintaining strict admissibility. To our knowledge, this is the first machine-learned heuristic guaranteed to be admissible.
2026-06-02
arXiv (prépublication)
doi.org
arxiv.org