Publications

Wagg: Cost-aware Aggregation of Windowing Operators in Stream Processing
Pritish Mishra
Ruoyu Deng
Alexandre da Silva Veith
Oana Balmau
Eyal de Lara
2026-06-19
International Conference on Mobile Systems, Applications and Services (published)
doi.org
Accelerated and Stable Convergence with Anchored Optimistic Method
Motahareh Sohrabi
Jianxin You
Simon Lacoste-Julien
Eduard Gorbunov
Gauthier Gidel
We study first-order methods for solving monotone variational inequalities arising in min-max optimization. Classical approaches such as the… (see more) extragradient method rely on two gradient queries per iteration, which limits their analysis and applicability in the online and stochastic settings. We propose a family of Generalized Optimistic Methods with Anchoring (GOMA), which combine two-time-scale optimistic updates with an anchoring term inspired by Halpern iteration. In the deterministic setting, GOMA achieves the optimal accelerated last-iterate rate
2026-06-18
arXiv (preprint)
doi.org
arxiv.org
Continual Test-Time Adaptation in Computer Vision: Methods, Benchmarks, and Future Directions
Sarthak Kumar Maharana
Shambhavi Mishra
Yunbei Zhang
shuaicheng niu
Taki Hasan Rafi
Jihun Hamm
Marco Pedersoli
Jose Dolz
Yunhui Guo
Deep neural nets achieve remarkable performance when training and test data share the same distribution, but this assumption frequently brea… (see more)ks in real-world deployment, where data undergoes continual distributional shifts. Continual Test-Time Adaptation (CTTA) addresses this challenge by adapting pretrained models to non-stationary target distributions on-the-fly, without access to source data or labeled targets, while mitigating two critical failure modes: catastrophic forgetting of source knowledge and error accumulation from noisy pseudo-labels over extended time horizons. In this comprehensive survey, we formally define the CTTA problem, analyze the diverse continual domain shift patterns that characterize different evaluation protocols, and propose a hierarchical taxonomy that categorizes existing methods into three families: optimization-based strategies (entropy minimization, pseudo-labeling, parameter restoration), parameter-efficient methods (normalization layer adaptation, adaptive parameter selection), and architecture-based approaches (teacher-student frameworks, adapters, visual prompting, masked modeling). We systematically review representative methods within each category and present comparative benchmarks and experimental results across standard evaluation settings. Finally, we discuss limitations of current approaches and highlight emerging research directions, including adaptation of foundation models and black-box systems, providing a roadmap for future research in robust continual test-time adaptation. We encourage visiting our repository at https://github.com/sarthaxxxxx/Awesome-Continual-Test-Time-Adaptation.
2026-06-18
Zenodo (preprint)
doi.org
Geometric Path Following for Autonomous Dynamic Soaring
Zihao Zhuo
Meyer Nahon
Inna Sharf
Autonomous dynamic soaring can be used to increase the endurance and range of unmanned aerial vehicles by harvesting energy from the vertica… (see more)l gradient of the horizontal wind. This study aims to develop a guidance and control strategy that allows precise following of an optimal dynamic-soaring path for a glider vehicle. The proposed control architecture combines a geometric path-following guidance law with an SO(3)-based attitude control law. High-fidelity six-degree-of-freedom simulation shows that the proposed method can achieve a position accuracy of 0.1 m for a glider with a wingspan of 2 m while adhering to the constraints present on a glider airframe. The high tracking accuracy makes it possible to conduct autonomous dynamic-soaring operations with patterns that were considered impossible in previous studies, such as a travel pattern mimicking the albatrosses’ dynamic-soaring pattern in close proximity to the ocean surface.
2026-06-18
Journal of Guidance, Control, and Dynamics (published)
doi.org
PrivacyAlign: Contextual Privacy Alignment for LLM Agents
Manveer Singh Tamber
Abhay Puri
Marc-Etienne Brunet
Perouz Taslakian
Jimmy Lin
Spandana Gella
AI agents acting on behalf of users are constantly making decisions, and for users to trust their agents, those decisions must align with wh… (see more)at they actually want. Privacy is an important alignment problem for agents: every message, post, or tool call an agent makes is a contextual judgment about what is appropriate to share, with whom, and under which conditions. Because such judgments depend on social expectations and norms, human judgment does not merely label privacy violations but also helps define them. While existing work relies on unreliable proxies for both training and evaluation, we place human judgment at the center of agentic privacy alignment. We introduce PrivacyAlign, a dataset of 1,350 samples with 3,516 detailed annotations from 599 unique annotators across diverse scenarios where current LLMs actually leak, and use it to ground both alignment training and automated evaluation in human privacy norms. Building on these annotations, we first show that conditioning LLM judges on human annotations and explanations for reference responses to the same prompt makes their judgments more reliable. We then introduce annotation-conditioned reward modeling, which uses these annotations to score new responses during RL, and show that small open-weight agents trained with this reward better align with human privacy norms, with strong gains on PrivacyAlign and existing privacy benchmarks for agents.
2026-06-18
arXiv (preprint)
doi.org
arxiv.org
The digital heartbeat: a qualitative descriptive study on women's views on preventing cardiovascular disease in primary care
Ilhem Chaima Bousbiat
Samira Abbasgholizadeh Rahimi
Roland Grad
Charo Rodriguez
BACKGROUND: This empirical study aims to explore women's perspectives on cardiovascular disease and the use of digital health interventions … (see more)(DHIs) for their primary prevention and to gather insights on essential features for developing artificial intelligent-enabled technologies. METHODS: Adopting a qualitative descriptive research design, we conducted 15 semi-structured, in-depth interviews via Zoom with women at higher risk for cardiovascular disease. Participants were women over 40 years old, residing in Quebec, with at least one cardiovascular disease risk factor, and proficient in English. Recruitment was from a McGill University-affiliated clinic. An inductive thematic analysis approach was used for data analysis. RESULTS: Five major themes were identified: (i) understanding cardiovascular disease in a variety of ways, (ii) barriers and challenges to preventing cardiovascular disease in women, (iii) women taking charge of their cardiovascular well-being, (iv) mixed perspectives regarding artificial intelligent-enabled technologies for cardiovascular disease prevention such as Xi-Care, and (v) range of suggestions for the format and design of a prospective artificial intelligent-enabled technologies. CONCLUSIONS: Despite the prevalence of cardiovascular disease, there is a significant knowledge gap among women regarding the chronic nature and manifestations of these diseases. Artificial intelligent-enabled technologies like Xi-Care, with the potential for customization and interactive engagement, could enhance the primary prevention of cardiovascular disease in women, providing valuable insights for the subsequent phases of the project leading to Xi-Care's development.
2026-06-18
Family Practice (published)
doi.org
FlowMaps: Modeling Long-Term Multimodal Object Dynamics with Flow Matching
Francesco Argenziano
Miguel Saavedra-Ruiz
Sacha Morin
Charlie Gauthier
Daniele Nardi
Liam Paull
Joint spatial and temporal understanding of 3D scenes is a crucial requirement for robots deployed in everyday household environments. Such … (see more)agents must not only comprehend and navigate spatial layouts, but also reason about how these spaces evolve over time. In particular, humans interact with objects daily, causing them to change position throughout the environment and making it difficult for robots to reliably associate current observations with previously seen objects. However, these interactions are not random: human habits and routines induce spatio-temporally consistent patterns in object locations, which robotic agents can potentially learn and then exploit for downstream tasks such as navigation. To this end, we introduce FlowMaps, a latent flow matching model for estimating multimodal distributions over the future locations of dynamic objects in a continuous 3D space. By learning the implicit dependencies among objects and their temporal evolution, FlowMaps predicts likely changes in object locations conditioned on past human interactions, while supporting generalization across previously unseen environments that share similar object routines. To demonstrate the utility of this method, we deploy FlowMaps in a downstream dynamic Object Navigation task in both simulated and real-world environments. Across more than 600 episodes, FlowMaps outperforms state-of-the-art approaches, showing that modeling object dynamics through continuous, multimodal spatio-temporal distributions improves robotic search and navigation in changing household environments. Code and additional material is available at https://fra-tsuna.github.io/flowmaps/.
2026-06-17
arXiv (preprint)
doi.org
arxiv.org
Hidden in Thought: Transferable Chain-of-Thought Artifacts Induce Harmful Behavior
Ali Khalil
Aly M. Kassem
Mohamed Abdelrazek
Santu Rana
Negar Rostamzadeh
Golnoosh Farnadi
We investigate whether harmful chain-of-thought (CoT) traces from compromised language models can transfer unsafe behaviour and be distilled… (see more) into reusable jailbreak attacks. Using an emergent-misalignment organism and a refusal-ablated jailbroken organism, we transplant harmful CoTs into
2026-06-17
arXiv (preprint)
doi.org
arxiv.org
Timely Availability and Accessibility of Health Data: Meeting the Challenge of Pan-Canadian Health Charter Principle 6
Kimberlyn McGrail
David L. Buckeridge
Pan-Canadian Health Data Charter Principle 6 envisions health data that are timely, accessible, meaningful, and comprehensive. This paper ex… (see more)amines three dimensions of this principle: what is envisioned for health data, for whom, and for what purposes. Comprehensive data include publicly funded care, privately paid services, patient-reported outcomes, and social determinants. To be meaningful, data require standardization and quality assurance. Timely access ranges from real-time clinical use to dependable research access. Accessibility depends on interoperable systems and secure environments. Current practices fall short across all dimensions, with fragmented, incomplete, and delayed data limiting timely access and effective use. Access and coverage remain uneven, reflecting structural and policy barriers. Despite these challenges, there are promising initiatives such as efforts to enhance national data stewardship. Advancing Principle 6 will require involvement of all interests in health data, including data stewards, policy-makers, providers, data users, and members of the public.
2026-06-17
Healthcare Management Forum (published)
doi.org
A call to integrate animal movement into biodiversity indicators
Ruth Y. Oliver
Katherine Hébert
Lacey F. Hughey
Luca Börger
Francesca Cagnacci
Nathan W. Cooper
Sarah C. Davidson
Andrew Gonzalez
Autumn‐Lynn Harrison
Jessica M. Kendall-Bar
Katie L. Millette
Joanna Mills Flemming
Thomas Mueller
Will Rogers
Talia Speaker
Jared A. Stabach
Marlee A. Tucker
Wenjing Xu
Scott W. Yanco
Briana Abrahms … (see 35 more)
Sara Beery
Roxanne S. Beltran
Lily K. Bentley
Larissa T. Beumer
Mary E. Bowers
Steven W. J. Canty
Ying‐Chi Chan
Juliet Cohen
Grant M. Connette
Eduardo Cuevas
Tammy E. Davies
Daniel C. Dunn
Diego Ellis‐Soto
Antonio Ferraz
John Fieberg
Kimberly R. Hall
Neil Hammerschlag
Anne G. Hertel
Dongmin Kim
Samara Manzin
Clive R. McMahon
Robin Naidoo
Aidin Niamir
A. Justin Nowakowski
Matthew B. Ogburn
Jonathan Pye
José Manuel Reyes‐González
Nicholas J. Russo
Christian Rutz
Amy L. Scarpignato
Stella F. Uiterwaal
Raqib Valli
Alessandra Vidal Meza
George Wittemyer
Laura J. Pollock
2026-06-16
Nature Reviews Biodiversity (published)
doi.org
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
Guillaume Lajoie
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)
doi.org
arxiv.org
Convex training of Lipschitz-regularized shallow neural networks
Chao Yin
Antoine Lesage-Landry
In this work, we introduce a training procedure for shallow neural networks that promotes robustness against adversarial attacks. We solve a… (see more) non-convex Lipschitz-regularized training program by introducing a convex restriction that can be efficiently solved to global optimality. Our approach can be employed as a post-processing step by taking a pre-trained network as an initial solution to then solving the convex program whose optimal network is guaranteed to be no worse than the initial one. We illustrate the improvements of our training procedure with experiments using real world datasets for regression tasks under an adversarial setting. We show numerically that solving our proposed convex program yields networks with lower objective values on the Lipschitz-regularized program compared to existing methods. Additionally, we show that on certain datasets, networks obtained using our convex training program are both more accurate and robust with respect to adversarial attacks.
2026-06-16
arXiv (preprint)
doi.org
arxiv.org