Publications

A unified 12-lead ECG-language model for interpretation and clinical-endpoint prediction
Juliette Beaulieu
Nicolas Dostie
B Mondesert
Ram Ahuja
Gilbert Jabbour
Shreya Shree Srikanth
Guillaume Marquis‐Gravel
Olivier Tastet
A. Sowa
Julia Cadrin‐Tourigny
Jacques Delfrate
Robert Avram
Automated electrocardiogram (ECG) interpretation has advanced, yet most systems remain narrow classifiers that emit fixed labels rather than… (see more) the narratives or endpoint-specific answers clinicians need. Generative approaches could instead produce rich narratives, but are constrained by the gap between continuous biosignals and discrete language tokens. Here we present DeepECG-Tok, which reframes ECG interpretation as a unified instruction-following problem. A residual vector-quantization tokenizer (QINCo) maps 12-lead waveforms to language-model-compatible tokens. Its frozen embeddings achieved a macro-averaged area under the receiver operating characteristic curve (AUROC) of 0.96 for 77-condition classification, outperforming supervised and self-supervised baselines. Aligned with a large language model, a single instruction-tuned model performed ECG interpretation, structured reporting and clinical endpoint prediction, including left ventricular ejection fraction, structural heart disease and atrial fibrillation risk, using 7.27 million question–answer pairs. Frozen-tokenizer diagnostic classification transferred without retraining to two external cohorts, retaining macro-averaged AUROCs of 0.88–0.90, and clinical-endpoint prediction transferred to two further cohorts (external LVEF ≤40% AUROC 0.74–0.76). Evaluated end to end using an ontology-grounded large language model as a judge, which achieved a mean agreement (Cohen’s κ) of 0.82 against two cardiologists, the unified instruction-tuned model scored 0.71 internally and 0.50–0.53 in the same external cohorts. In a blinded reader study, board-certified cardiologists and residents rated its free-text reports comparably to reference clinician reports, with a paired win–tie–loss distribution of 32:35:33 and a forced-choice preference of 0.53 among decided cases, with no significant difference between the model and reference reports in either comparison. These findings establish discrete ECG tokenization as a foundation for general-purpose models that generate clinically useful interpretations and answer diverse questions directly from cardiac waveforms.
Controllable and Content-Based Recommendations
Traditional recommendation systems rely on latent (dense) representations, making them difficult to interpret and control. We propose the Co… (see more)ntrollable and Content-Based Recommendations (CCBR) framework, which builds its recommendations from textual user profile representations. CCBR plugs into collaborative filtering models and introduces controllability via text bottlenecks. We show that CCBR enables text-based and multimodal interventions, allowing users to steer the model towards the directions they prefer. Different from existing controllable recommendation systems, CCBR infers the text summaries directly from item contents (images, audio or video). Across image-, audio-, and video-based datasets, we demonstrate that the proposed framework obtains competitive model performance with standard (latent-representation) models while providing controllable model summaries via text. The model also outperforms TEARS, a recent baseline for controllable recommendation systems. Through systematic interventions, we demonstrate the efficacy of the user steering mechanism.
Cortical microstructural integrity predicts an exploitation bias in older adulthood
Patrick Hewan
Alfie Wearn
Jeremy Hogeveen
Kayla Williams
R Nathan Spreng
Gary R. Turner
Sylvia Villeneuve
Judes Poirier
John C S Breitner
Andrée-Ann Baril
Bellec Pierre
Véronique Bohbot
Mallar Chakravarty
D Louis Collins
Mahsa Dadar
Simon Ducharme
Alan Evans
Claudine Gauthier … (see 80 more)
Maiya R Geddes
Rick Hoge
Yasser Ituria‐Medina
Gerhard Multhaup
Lisa-Marie Münter
Natasha Rajah
Pedro Rosa-Neto
Taylor Schmitz
Soucy Jp
Nathan Spreng
Christine Tardif
Etienne Vachon-Presseau
Mohammadali Javanray
Meishan Ai
Philippe Amouyel
Nicholas Ashton
Gabriel Aumont‐Rodrigue
Julie Bailly
Guilia Baracchini
Kaj Blennow
Christian Bocti
Lianne Boisvert
Sophie Boutin
Ann Brinkmalm Westman
A P Dagher
Xing Dai
Samir Das
Marina Dauar‐Tedeschi
Louis De Beaumont
Christine Déry
Maxime Descoteaux
Elena Drobotea
M Elie
Alfonso Fajardo Valdez
Vladimir Fonov
David Morgan
Jonathan Gallago
Greco Cr
Louise Hudon
Gabriel Jean
Anne Labonté
Robert Laforce
Marc Lalancette
Jean-Charles Lambert
Jeannie‐Marie Leoutsakos
Danaé Lussier Dumouchel
B Misic
Béry Mohammediyan
Holly NewboldFox
Eugenia Nita Capota
Alix Noly‐Gandon
Adrian Eduardo Noriega de la Colina
Pierre Orban
Valentin Ourry
Cynthia Picard
Alexa Pichet Binette
A. L. Poirier
Nathalie Prenevost
Ting Qiu
Marc James Quesnel
Charles Ramassamy
Jean‐Michel Raoult
Jordana Remz
Safa Sanami
Frederic St‐Onge
Cherie Strikwerda‐Brown
Elisabeth Sylvain
Andràs Tikàsz
Christina Tremblay
Stefanie Tremblay
Jennifer Tremblay‐Mercier
Stéphanie Tullo
Irem Ulku
Paolo Vitali
Yara Yakoub
Robert Zatorre
Henrik Zetterberg
Pierre Bellec
Jean-Paul Soucy
Claudia Greco
OBJECTIVES: Prefrontal regions are implicated in explore-exploit decision-making during foraging. Older adults often show an exploitation bi… (see more)as, and this age period is also marked by deteriorating prefrontal myelination. To investigate whether these phenomena are linked, we examined whether lower magnetization transfer saturation (MTsat), a myelin-sensitive quantitative MRI (qMRI) measure, in these regions predicts greater exploitation bias during foraging, and whether cortical microstructure is a better predictor of bias than macrostructure (i.e., cortical thickness). METHODS: Cognitively healthy older adults with familial risk of Alzheimer's disease (AD) (N=118, 60-88 years) completed a foraging task indexing explore-exploit decision-making. qMRI was used to derive MTsat values for the frontopolar cortex (FPC), medial orbitofrontal cortex (OFC), rostral middle frontal gyrus (rMFG), dorsal anterior cingulate cortex (dACC), as well as the locus coeruleus (LC), a core subcortical region strongly implicated in explore-exploit decision-making. Secondary analyses examined associations between available AD risk markers and foraging. RESULTS: Lower MTsat in the FPC, OFC, rMFG, and LC was associated with an exploitation bias, with LC and FPC emerging as the strongest predictors. No relationship was observed for the dACC. MTsat remained a significant predictor of foraging after controlling for cortical thickness. Observed associations were largely unrelated to AD risk markers. DISCUSSION: Individual differences in cortical microstructural integrity within a well-defined explore-exploit circuit are associated with an exploitative decision-making bias in older adults. These findings highlight the value of qMRI microstructural integrity markers, beyond standard macrostructural assays, in characterizing the neural correlates of exploitation biases in later life.
FedAgentKE: Federated Semantic Knowledge Evolution for Heterogeneous Agents
Weihao Li
Ziyang Song
Large language model (LLM)-based agents increasingly rely on reasoning, tool use, and iterative execution, yet existing agent frameworks sti… (see more)ll operate largely in isolation. While recent memory-based agent systems improve individual agents through local retrieval and workflow reuse, local experiences remain fragmented across isolated agent frameworks, limiting cross-framework knowledge transfer and collaborative reasoning evolution. We propose FedAgentKE, a lightweight framework for Federated Semantic Knowledge Evolution across heterogeneous agents. FedAgentKE enables distributed agent frameworks to collaboratively evolve transferable reasoning abstractions through iterative semantic knowledge distillation, aggregation, and adaptation without sharing raw reasoning trajectories. Experiments demonstrate consistent improvements under both cross-framework and cross-task settings, highlighting the potential of federated semantic knowledge evolution for future collaborative agent ecosystems.
High-resolution dissection of concept acquisition in different families of protein language models
Robert M. Vernon
Christopher J. Langmead
Protein language models have been increasingly successful on tasks ranging from fitness prediction to functional design, yet what biological… (see more) knowledge they acquire and where it is encoded within their internal representations remain underexplored. Through a high-resolution layer-by-layer interpretability analysis of 8 models from the ESM2 and AMPLIFY families on 22 concepts from human proteome annotations, we found that these models encode concepts of increasing levels of complexity along their depth: basic physicochemical properties and linear motifs are best captured by early-layer embeddings, secondary structure from subsequent layers, and domain-level semantics from middle layers. Principal component projections of these embeddings showed that they separate biologically meaningful protein groupings, and molecular-biology-inspired interventions demonstrated that pLM embeddings can discriminate phosphomimic-active from inactive mutants. Perhaps surprisingly, we observed that pretraining data and compute had a greater impact on the linear emergence of biological concepts than scaling up parameters. By revealing where biological knowledge is captured in pLMs and which choices shape its emergence, our work offers insights to develop more robust, biologically grounded protein language models.
How Do AI Coding Agents Contribute to Software Development? an Empirical Study of Agentic Pull Requests
Iren Mazloomzadeh
Mohammad Mehdi Morovati
Recent advances in large language models and their rapid adoption across software engineering tasks have made Artificial Intelligence (AI) c… (see more)oding agents an integral component of modern software development workflows. While developers increasingly benefit from these coding agents, their impact on software quality remains insufficiently understood. In particular, how agentic contributions evolve across the software development lifecycle has not been thoroughly investigated. This study aims to characterize agentic pull requests (PR) in comparison to human generated PRs and to examine how their properties change across different stages of the development lifecycle. Using the AIDev dataset, we first analyze how differences in merge rates between agentic and human generated PRs vary over time. We then identify the types of development tasks where AI coding agents are predominantly applied and investigate how these task distributions evolve across development quarters. Finally, we compare a set of key characteristics of agentic and human generated PRs, focusing on their implications for software quality and their temporal dynamics. Overall, our findings provide an empirical and longitudinal perspective on the role of AI coding agents in software development, offering a more nuanced understanding of their benefits and limitations in real-world practices.
MSpecTmol: A Multi-Modal Spectroscopic Learning Framework for Automated Molecular Structure Elucidation
Wenjie Du
Xiaohan Qin
Ye Wei
Jun Xia
Spectroscopic techniques are indispensable for the elucidation of molecular structures, particularly for novel molecules with unknown config… (see more)urations. However, a fundamental limitation of any single spectroscopic modality is that it provides an inherently circumscribed and fragmented view, capturing only specific facets of the complete molecular structure, which is often insufficient for unequivocal and robust characterization. Consequently, the integration of data from multiple spectroscopic sources is imperative to overcome these intrinsic limitations and achieve a comprehensive and accurate structural characterization. In this work, we introduce \textbf{MSpecTmol}, a novel \textbf{M}ulti-modal \textbf{Spec}trum information fusion learning framework for automated \textbf{Mol}ecular structure elucidation. By extending information bottleneck theory, our framework provides a principled and adaptive approach to fusing spectra. It designates a primary modality to extract core molecular features while leveraging auxiliary inputs to enrich the representation. To validate the end-to-end effectiveness of our framework, we design a two-fold evaluation: molecular substructure classification to probe its discriminative power in identifying substructures, and extends this knowledge to reconstruct plausible 3D structures. Our results not only demonstrate state-of-the-art performance in molecular substructure classification but also achieve near-experimental accuracy (\textasciitilde 0.68\AA) in molecular conformation reconstruction. These findings underscore the model’s capacity to learn interpretable features aligned with chemical intuition, thereby paving the way for future advances in automated and reliable spectroscopic analysis. Our code can be found at \href{https://anonymous.4open.science/r/MspecTmol-6B4D}{https://anonymous.4open.science.}
Overview of FinMMEval 2026 Task 1: Multilingual Financial Multiple-Choice Question Answering
Zhuohan Xie
Yuyang Dai
Rania Elbadry
Vanshikaa Jani
Georgi Georgiev
Dimitar Dimitrov
Fan Zhang
Xueqing Peng
Lingfei Qian
Jimin Huang
Jiahui Geng
Yankai Chen
Yuxia Wang
Ivan Koychev
Veselin Stoyanov
Mingzi Song
Yu Chen
Xue Liu … (see 1 more)
Preslav Nakov
FinMMEval 2026 Task 1 evaluates multilingual financial multiple-choice question answering in English, Chinese, Arabic, and Hindi. The task t… (see more)ests whether systems can select the correct answer to finance questions involving domain terminology, numerical interpretation, and conceptual financial reasoning across languages and scripts. The final-test set contains 800 questions, with 200 questions per language; gold answers were withheld during submission, and each language was ranked independently by accuracy. The final leaderboards contain 13 English, 11 Chinese, 11 Arabic, and 10 Hindi ranked submissions. Top accuracies range from 92.0% in Hindi to 97.5% in English and Arabic, with the same leading teams appearing near the top across all four languages. The documented systems used retrieval augmentation, direct answer-option scoring, language-specific prompting, selective self-consistency, confidence checks, and LLM-based review stages.
Overview of FinMMEval 2026 Task 2: Multilingual Financial Short-Answer Question Answering
Zhuohan Xie
Xueqing Peng
Georgi Georgiev
Dimitar Dimitrov
Yuyang Dai
Rania Elbadry
Vanshikaa Jani
Lingfei Qian
Fan Zhang
Jimin Huang
Jiahui Geng
Yankai Chen
Yuxia Wang
Ivan Koychev
Veselin Stoyanov
Mingzi Song
Yu Chen
Xue Liu … (see 1 more)
Preslav Nakov
FinMMEval 2026 Task 2 evaluates short-answer financial question answering over multilingual evidence. Each final-test item pairs an English … (see more)question with financial statements and news in English, Chinese, Japanese, Spanish, and Greek. Participating systems submit one concise answer per item in JSONL format. The final-test set contains 256 items, split evenly between easy and expert tiers; each tier contains four question templates instantiated over 32 company-report groups. Gold answers were withheld during submission, and systems were ranked by macro-averaged item-level ROUGE-1 F1 against organizer-held reference answers. The final leaderboard includes 12 ranked submissions. The strongest systems are closely clustered, with the top four separated by less than one percentage point in ROUGE-1 F1. The submitted system papers document retrieval-augmented generation, cross-lingual evidence handling, structured prompting, answer compression, and validation strategies.
SpatialJEPA: JEPA-inspired graph-context distillation for spatially aware multiomics integration
Abstract Computational frameworks for integrating spatial genomics modalities extend cell-based representation learning across molecular lay… (see more)ers, but many paired RNA–ATAC datasets are dissociated and lack spatial coordinates. We introduce SpatialJEPA , a JEPA-inspired teacher–student framework for transferring spatial context from spatial multiomics data to non-spatial multiome data. In contrast to patch- or feature-masking objectives, SpatialJEPA masks spatial context by replacing the teacher’s spatial neighborhood graph with a self-only identity graph during student training, making the spatial sample appear dissociated to the student. The student learns to match teacher embeddings from this graph-context-restricted view and can therefore be applied to dissociated RNA–ATAC data at inference time. In mouse brain multiomics, the resulting representation supports source–target alignment, recovers spatially organized transcriptomic and chromatin-accessibility programs, and shows concordance with ligand–receptor pathway structure compared with non-spatial references.
AIMSDistill: Distilling knowledge from specialised AI teachers for cross-jurisdictional compliance analysis of modern slavery statements
Adriana Eufrosina Bora
Duoyi Zhang
Md. Abul Bashar
Richi Nayak
Kerrie Mengersen
Modern slavery in global supply chains remains a critical concern, prompting governments in countries such as the UK, Australia, and Canada … (see more)to introduce corporate transparency legislation, including Modern Slavery Acts. However, the growing volume of compliance reports has made manual analysis increasingly impractical. This challenge is compounded by the fact that many government agencies and NGOs responsible for assessing these statements operate in resource-constrained environments, limiting their ability to deploy large-scale language models. To address both issues, we present a novel AI framework for cross-jurisdictional compliance analysis. Our approach integrates four specialised teacher models, each trained using contrastive learning, prompt engineering, pseudo-labelling, and context-enhanced learning, which are distilled into a single lightweight, computationally efficient student model. This design enhances generalisability across legal frameworks while remaining affordable to deploy in low-resource settings. Experimental evaluations demonstrate that our method achieves higher accuracy and efficiency than existing approaches, offering a scalable, accessible solution for regulators and other stakeholders.
Enhancing Psychiatry Training Using an Agentic AI Simulated Consultation Tool: Prospective Cohort Study
Alice Rueda
Huda F Al-Shamali
Zack Cote
Niloy Roy
Reinhard Janssen‐Aguilar
Jithin Joseph
Bazen Gashaw Teferra
Mohammad Amin Kamaleddin
Lisa Burback
Olga Winkler
Bill Kapralos
Andrei Torres
Sridhar Krishnan
Andrew J. Greenshaw
Alexandre Hudon
Sanjeev Sockalingam
Yanbo Zhang
Adam Dubrowski … (see 1 more)
Venkat Bhat
Background: Canadian psychiatry residents must demonstrate consultation competency, assessed using the standardized assessment of a clinical… (see more) encounter report (STACER). However, opportunities to practice these skills and receive constructive assessment remain limited in clinical settings. Objective: This study aimed to evaluate the technical feasibility of an agentic AI system designed to support psychiatry residents' consultation competence through simulated patient encounters with a patient agent and structured feedback from a rater agent. Methods: We conducted a two-phase technical feasibility prospective single-arm cohort study of the STACER Agentic System, a large language model-based platform integrating a patient agent and a rater agent. Phase 1 involved automated evaluation of the patient agent using a psychiatrist agent across 227 synthetic major depressive disorder cases. Performance was assessed using DeepEval metrics (correctness, clarity, medical faithfulness, turn relevance, and role adherence) with descriptive statistics and 95% CIs. Phase 2 involved a preliminary user study with 14 convenience-sampled participants: a total of 5 members of the clinical research team and 9 psychiatry residents from the University of Alberta. Participants completed simulated diagnostic interviews and case presentations. Performance was evaluated using STACER-based scoring by the rater agent and 2 psychiatrists. Interrater reliability was assessed using intraclass correlation coefficients (α=.05). Participants rated realism, behavioral consistency, psychiatric nuance, and feedback utility using Likert scales and free-text answers. Results: The patient agent demonstrated high behavioral (51/56, 91.07%) and symptom fidelity (105/110, 95.45%), with strong automated performance (medical faithfulness mean 0.99, 95% CI 0.99-1.00; turn relevance 0.99, 95% CI 0.986-0.992). Participants rated simulations as psychiatrically plausible and diagnostically useful, particularly for depressive symptom representation, although rapport building was moderate (mean 2.78, SD 1.56 to mean 3.00, SD 1.41, out of 5.00) due to limited nonverbal cues. The rater agent generated structured STACER-aligned feedback with high intrarater consistency, especially at the section subtotal level. Interrater reliability with psychiatrists was poor at the item level (intraclass correlation coefficient range=0.25-0.49) but improved to good-to-excellent agreement at the section level for psychiatry resident sessions (intraclass correlation coefficient range=0.89-0.93). The rater agent's scores fell between those of the 2 psychiatrists for the clinical research team and were lower than both human raters for psychiatry residents. Conclusions: The STACER Agentic System demonstrates the technical feasibility of using agentic AI to simulate psychiatric consultations and deliver STACER-aligned formative feedback. By combining adaptive multiturn psychiatric simulation with competency-based evaluation, it shows promise in supporting cognitive aspects of consultation, though it remains limited in facilitating relational skills such as rapport building. These findings suggest agentic AI could expand scalable, low-risk opportunities for deliberate practice and formative feedback in competency-based psychiatric education. Further controlled studies are needed to evaluate educational effectiveness and integration into residency training.