Publications

Multi-Agent AI Framework for Threat Mitigation and Resilience in Machine Learning Systems
Armstrong Foundjem
Lionel Nganyewou Tidjon
Leuson Da Silva
Foutse Khomh
Machine learning (ML) increasingly underpins foundation models and autonomous pipelines in high-stakes domains such as finance, healthcare, … (voir plus)and national infrastructure, rendering these systems prime targets for sophisticated adversarial threats. Attackers now leverage advanced Tactics, Techniques, and Procedures (TTPs) spanning data poisoning, model extraction, prompt injection, automated jailbreaking, training data exfiltration, and—more recently—preference-guided black-box optimization that exploits models’ own comparative judgments to craft successful attacks iteratively. These emerging text-only, query-based methods demonstrate that larger and better-calibrated models can be paradoxically more vulnerable to introspection-driven jailbreaks and cross-modal manipulations. While traditional cybersecurity frameworks offer partial mitigation, they lack ML-specific threat modeling and fail to capture evolving attack vectors across foundation, multimodal, and federated settings. Objective: This research empirically characterizes modern ML security risks by identifying dominant attacker TTPs, exposed vulnerabilities, and lifecycle stages most frequently targeted in foundation-model, multimodal, and retrieval-augmented (RAG) pipelines. The study also assesses the scalability of current defenses against generative and introspection-based attacks, highlighting the need for adaptive, ML-aware security mechanisms. Methods: We conduct a large-scale empirical analysis of ML security, extracting 93 distinct threats from multiple sources: real-world incidents in MITRE ATLAS (26), the AI Incident Database (12), and peer-reviewed literature (55), supplemented by 854 ML repositories from GitHub and the Python Advisory database. A multi-agent reasoning system with enhanced Retrieval-Augmented Generation (RAG)—powered by ChatGPT-4o (temperature 0.4)—automatically extracts TTPs, vulnerabilities, and lifecycle stages from over 300 scientific articles using evidence-grounded reasoning. The resulting ontology-driven threat graph supports cross-source validation and lifecycle mapping. Results: Our analysis uncovers multiple unreported threats beyond current ATLAS coverage, including model-stealing attacks against commercial LLM APIs, data leakage through parameter memorization, and preference-guided query optimization enabling text-only jailbreaks and multimodal adversarial examples. Gradient-based obstinate attacks, MASTERKEY automated jailbreaking, federated learning poisoning, diffusion backdoor embedding, and preference-oriented optimization leakage emerge as dominant TTPs, disproportionately impacting pretraining and inference. Graph-based dependency analysis shows that specific ML libraries and model hubs exhibit dense vulnerability clusters lacking effective issue-tracking and patch-propagation mechanisms. Conclusion: This study underscores the urgent need for adaptive, ML-specific security frameworks that address introspection-based and preference-guided attacks alongside classical adversarial vectors. Robust dependency management, automated threat intelligence, and continuous monitoring are essential to mitigate supply-chain and inference-time risks throughout the ML lifecycle. By unifying empirical evidence from incidents, literature, and repositories, this research delivers a comprehensive threat landscape for next-generation AI systems and establishes a foundation for proactive, multi-agent security governance in the era of large-scale and generative AI.
2026-01-18
ACM Transactions on Software Engineering and Methodology (publié)
doi.org
Toward Faithful Explanations in Acoustic Anomaly Detection
Maab Elrashid
Anthony Deschênes
Yusuf Cem Sübakan
Mirco Ravanaelli
Rémi Georges
Michael Morin
Interpretability is essential for user trust in real-world anomaly detection applications. However, deep learning models, despite their stro… (voir plus)ng performance, often lack transparency. In this work, we study the interpretability of autoencoder-based models for audio anomaly detection, by comparing a standard autoencoder (AE) with a mask autoencoder (MAE) in terms of detection performance and interpretability. We applied several attribution methods, including error maps, saliency maps, SmoothGrad, Integrated Gradients, GradSHAP, and Grad-CAM. Although MAE shows a slightly lower detection, it consistently provides more faithful and temporally precise explanations, suggesting a better alignment with true anomalies. To assess the relevance of the regions highlighted by the explanation method, we propose a perturbation-based faithfulness metric that replaces them with their reconstructions to simulate normal input. Our findings, based on experiments in a real industrial scenario, highlight the importance of incorporating interpretability into anomaly detection pipelines and show that masked training improves explanation quality without compromising performance.
2026-01-18
ArXiv (prépublication)
doi.org
arxiv.org
Press Start to Charge: Videogaming the Online Centralized Charging Scheduling Problem
Alireza Ghahtarani
Martin Cousineau
Amir-massoud Farahmand
Jorge E. Mendoza
2026-01-17
ArXiv (prépublication)
doi.org
arxiv.org
Quantification of head and neck cancer patients' anatomical changes during radiotherapy: Toward the prediction of replanning need
Odette Rios‐Ibacache
James Manalad
Kayla O'Sullivan‐Steben
Emily Poon
Luc Galarneau
Julia Khriguian
George Shenouda
J. Kildea
Abstract Background Head and neck cancer (HNC) patients undergoing radiotherapy (RT) may experience anatomical changes during treatment, whi… (voir plus)ch can compromise the validity of the initial treatment plan, necessitating replanning. However, ad hoc replanning disrupts clinical workflows and increases workload. Currently, no standardized method exists to quantify anatomical variation that necessitates replanning. Purpose This project aimed to create geometrical metrics to describe anatomical changes in HNC patients during RT. The usefulness of these metrics was evaluated by a univariate analysis and through machine learning (ML) models to predict the need for replanning. Methods A cohort of 150 HNC patients treated at McGill University Health Centre was analyzed. Based on the shapes of the RT structures (body, PTV, mandible, neck, and submandibular contours), we developed 43 metrics and automatically calculated them through a Python pipeline that we called HNGeoNatomyX. Univariate analysis using linear regression was conducted to obtain the rate of change of each metric. We also obtained the relative variation of each metric between the pre‐treatment and replanning‐requested scans. Fraction‐specific ML models (incorporating information available up to and including the specific fraction) for fractions 5, 10, and 15 were built using metrics, clinical data, and feature selection techniques. Model performance was estimated with repeated stratified 5‐fold cross‐validation resampling technique and the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. Results Univariate analysis showed that body‐ and neck‐related metrics were most predictive of replanning need. Our best specific multivariate models for fractions 5, 10, and 15 yielded testing scores of 0.82, 0.70, and 0.79, respectively. Our models early predicted replanning for 76% of the true positives. Conclusions The created metrics have the potential to characterize and distinguish which patients will necessitate RT replanning. They show promise in guiding clinicians to evaluate RT replanning for HNC patients and streamline workflows.
2026-01-16
Journal of Applied Clinical Medical Physics (publié)
doi.org
arxiv.org
Manifold Learning for Olfactory Habituation to Strongly Fluctuating Backgrounds
François X. P. Bourassa
Paul François
Gautam Reddy
Massimo Vergassola
Animals rely on their sense of smell to survive, but important olfactory cues are mixed with confounding background odors that fluctuate due… (voir plus) to atmospheric turbulence. It is unclear how the olfactory system habituates to such stochastic backgrounds to detect behaviorally important odors. Here, we explicitly consider the high-dimensional nature of odor coding, the natural statistics of odor fluctuations, and the architecture of the early olfactory pathway. We show that their combination favors a manifold learning mechanism for olfactory habituation over alternatives based on predictive filtering. Manifold learning is implemented in our model by a biologically plausible network of inhibitory interneurons in the early olfactory pathway. We demonstrate that plasticity rules based on the Intrator, Bienenstock, Cooper, and Munro (IBCM) model or an online principal components analysis algorithm are effective at implementing this mechanism in turbulent conditions and outperform previous models relying on mean background subtraction. Interneurons with an IBCM plasticity rule acquire selectivity to independently varying odors. This manifold learning mechanism offers a path toward distinguishing plasticity rules in experiments and could be leveraged by other biological circuits facing fluctuating environments.
2026-01-15
PRX Life (publié)
doi.org
Same/Other/All K‐Fold Cross‐Validation for Estimating Similarity of Patterns in Data Subsets
Toby Dylan Hocking
Gabrielle Thibault
C. S. Bodine
Paul Nelson Arellano
Alexander F. Shenkin
Olivia Jasmine Lindly
2026-01-15
Statistical Analysis and Data Mining: An ASA Data Science Journal (publié)
doi.org
Discrete Feynman-Kac Correctors
Mohsin Hasan
Viktor Ohanesian
Artem Gazizov
Yoshua Bengio
Alán Aspuru-Guzik
Roberto Bondesan
Marta Skreta
Kirill Neklyudov
Discrete diffusion models have recently emerged as a promising alternative to the autoregressive approach for generating discrete sequences.… (voir plus) Sample generation via gradual denoising or demasking processes allows them to capture hierarchical non-sequential interdependencies in the data. These custom processes, however, do not assume a flexible control over the distribution of generated samples. We propose Discrete Feynman-Kac Correctors, a framework that allows for controlling the generated distribution of discrete masked diffusion models at inference time. We derive Sequential Monte Carlo (SMC) algorithms that, given a trained discrete diffusion model, control the temperature of the sampled distribution (i.e. perform annealing), sample from the product of marginals of several diffusion processes (e.g. differently conditioned processes), and sample from the product of the marginal with an external reward function, producing likely samples from the target distribution that also have high reward. Notably, our framework does not require any training of additional models or fine-tuning of the original model. We illustrate the utility of our framework in several applications including: efficient sampling from the annealed Boltzmann distribution of the Ising model, improving the performance of language models for code generation and amortized learning, as well as reward-tilted protein sequence generation.
2026-01-14
arXiv (prépublication)
doi.org
arxiv.org
Inference-time Physics Alignment of Video Generative Models with Latent World Models
Jianhao Yuan
Xiaofeng Zhang
Felix Friedrich
Nicolas Beltran-Velez
Melissa Hall
Reyhane Askari Hemmat
Xiaochuang Han
Nicolas Ballas
Michal Drozdzal
Adriana Romero
State-of-the-art video generative models produce promising visual content yet often violate basic physics principles, limiting their utility… (voir plus). While some attribute this deficiency to insufficient physics understanding from pre-training, we find that the shortfall in physics plausibility also stems from suboptimal inference strategies. We therefore introduce WMReward and treat improving physics plausibility of video generation as an inference-time alignment problem. In particular, we leverage the strong physics prior of a latent world model (here, VJEPA-2) as a reward to search and steer multiple candidate denoising trajectories, enabling scaling test-time compute for better generation performance. Empirically, our approach substantially improves physics plausibility across image-conditioned, multiframe-conditioned, and text-conditioned generation settings, with validation from human preference study. Notably, in the ICCV 2025 Perception Test PhysicsIQ Challenge, we achieve a final score of 62.64%, winning first place and outperforming the previous state of the art by 7.42%. Our work demonstrates the viability of using latent world models to improve physics plausibility of video generation, beyond this specific instantiation or parameterization.
2026-01-14
ArXiv (prépublication)
doi.org
arxiv.org
Multilinguality as Sense Adaptation
Jan Christian Blaise Cruz
David Ifeoluwa Adelani
Alham Fikri Aji
2026-01-14
ArXiv (prépublication)
doi.org
arxiv.org
BON in a Box: An Open and Collaborative Platform for Biodiversity Monitoring, Indicator Calculation, and Reporting
Jory Griffith
Jean-Michel Lord
Michael Catchen
María Isabel Arce-Plata
F Guillaume Blanchet
Mathusan Chandramohan
M Camila Diaz-Corzo
Gravel Dominique
César Gutiérrez
Isabelle S. Helfenstein
Sean Hoban
Jamie M. Kass
Linda Laikre
Guillaume Larocque
Deborah M. Leigh
Brian Leung
Alicia Mastretta‐Yanes
Katie L. Millette
Maria Alejandra Molina Berbeo
Dat Nguyen … (voir 27 de plus)
Kari Norman
María Helena Olaya-Rodríguez
Simon Pahls
Kaitlyn M. Pereira
Pedro R. Peres‐Neto
Timothée Poisot
Laura J. Pollock
Juan Carlos Rey-Velasco
Víctor J. Rincón-Parra
Claudia Roeoesli
François Rousseu
Lina María Sánchez-Clavijo
Meredith C. Schuman
Oliver Selmoni
Jessica M. da Silva
Erika Suárez-Valencia
Thilina D. Surasinghe
Eren Turak
Luis Urbina
Sarah Valentin
Noah Wightman
Juan Zuloaga
Maria Cecilia Londoño
Andrew Gonzalez
Manuel Galvez Bohorquez
Luis Felipe Verdeja González
Diego Moreno
Abstract The Convention on Biological Diversity’s Kunming–Montreal Global Biodiversity Framework (GBF) sets ambitious goals to protect a… (voir plus)nd restore biodiversity. It includes a monitoring framework that mandates countries to track progress toward these goals using indicators that summarize biodiversity trends. Calculating indicators is challenging for countries because of fragmented biodiversity monitoring efforts, technical barriers, a lack of available data and tools, and capacity bottlenecks. The BON in a Box platform for biodiversity monitoring and indicator calculation, developed by the Group on Earth Observations Biodiversity Observation Network, was created to address these challenges by providing open, transparent, and reproducible analysis pipelines that convert data into essential biodiversity variables and indicators. These pipelines are built by experts and contributed by the community, follow FAIR principles, and help scientists apply their research to coordinate biodiversity monitoring efforts, build capacity to track progress toward the GBF, and affect policy change.
2026-01-12
BioScience (publié)
doi.org
Evaluating Implicit Regulatory Compliance in LLM Tool Invocation via Logic-Guided Synthesis
Da Song
Yuheng Huang 0004
Boqi Chen
Tianshuo Cong
Randy Goebel
Lei Ma 0003
Foutse Khomh
The integration of large language models (LLMs) into autonomous agents has enabled complex tool use, yet in high-stakes domains, these syste… (voir plus)ms must strictly adhere to regulatory standards beyond simple functional correctness. However, existing benchmarks often overlook implicit regulatory compliance, thus failing to evaluate whether LLMs can autonomously enforce mandatory safety constraints. To fill this gap, we introduce LogiSafetyGen, a framework that converts unstructured regulations into Linear Temporal Logic oracles and employs logic-guided fuzzing to synthesize valid, safety-critical traces. Building on this framework, we construct LogiSafetyBench, a benchmark comprising 240 human-verified tasks that require LLMs to generate Python programs that satisfy both functional objectives and latent compliance rules. Evaluations of 13 state-of-the-art (SOTA) LLMs reveal that larger models, despite achieving better functional correctness, frequently prioritize task completion over safety, which results in non-compliant behavior.
2026-01-12
ArXiv (prépublication)
doi.org
arxiv.org
The oneirogen hypothesis: modeling the hallucinatory effects of classical psychedelics in terms of replay-dependent plasticity mechanisms
Colin Bredenberg
Fabrice Normandin
Blake Richards
Guillaume Lajoie
Abstract Classical psychedelics induce complex visual hallucinations in humans, generating percepts that are co-herent at a … (voir plus)low level, but which have surreal, dream-like qualities at a high level. While there are many hypotheses as to how classical psychedelics could induce these effects, there are no concrete mechanistic models that capture the variety of observed effects in humans, while remaining consistent with the known pharmacological effects of classical psychedelics on neural circuits. In this work, we propose the “oneirogen hypothesis”, which posits that the perceptual effects of classical psychedelics are a result of their pharmacological actions inducing neural activity states that truly are more similar to dream-like states. We simulate classical psychedelics’ effects via manipulating neural network models trained on perceptual tasks with the Wake-Sleep algorithm. This established machine learning algorithm leverages two activity phases, a perceptual phase (wake) where sensory inputs are encoded, and a generative phase (dream) where the network internally generates activity consistent with stimulus-evoked responses. We simulate the action of psychedelics by partially shifting the model to the ‘Sleep’ state, which entails a greater influence of top-down connections, in line with the impact of psychedelics on apical dendrites. The effects resulting from this manipulation capture a number of experimentally observed phenomena including the emergence of hallucinations, increases in stimulus-conditioned variability, and large increases in synaptic plasticity. We further provide a number of testable predictions which could be used to validate or invalidate our oneirogen hypothesis.
2026-01-12
eLife (accepté)
doi.org