Publications

TRUTH: Teaching LLMs to Rerank for Truth in Misinformation Detection
Hao Yu
Shenyang Huang
Zachary Yang
Maximilian Puelma Touzel
Kellin Pelrine
Jean-François Godbout
Reihaneh Rabbany
2025-07-24
colmweb.org/COLM/2025/Workshop/SoLaR (poster)
openreview.net
openreview.net
WiSE-OD: Benchmarking Robustness in Infrared Object Detection
Heitor Rapela Medeiros
Atif Belal
Masih Aminbeidokhti
Éric Granger
Marco Pedersoli
2025-07-24
ArXiv (prépublication)
doi.org
arxiv.org
Can LLMs Reason Abstractly Over Math Word Problems Without CoT? Disentangling Abstract Formulation From Arithmetic Computation
Ziling Cheng
Meng Cao
Leila Pishdad
Yanshuai Cao
Jackie CK Cheung
Final-answer-based metrics are commonly used for evaluating large language models (LLMs) on math word problems, often taken as proxies for r… (voir plus)easoning ability. However, such metrics conflate two distinct sub-skills: abstract formulation (capturing mathematical relationships using expressions) and arithmetic computation (executing the calculations). Through a disentangled evaluation on GSM8K and SVAMP, we find that the final-answer accuracy of Llama-3 and Qwen2.5 (1B-32B) without CoT is overwhelmingly bottlenecked by the arithmetic computation step and not by the abstract formulation step. Contrary to the common belief, we show that CoT primarily aids in computation, with limited impact on abstract formulation. Mechanistically, we show that these two skills are composed conjunctively even in a single forward pass without any reasoning steps via an abstract-then-compute mechanism: models first capture problem abstractions, then handle computation. Causal patching confirms these abstractions are present, transferable, composable, and precede computation. These behavioural and mechanistic findings highlight the need for disentangled evaluation to accurately assess LLM reasoning and to guide future improvements.
2025-07-23
colmweb.org/COLM/2025/Workshop/XLLM-Reason-Plan (publié)
openreview.net
openreview.net
Enhancing Logical Reasoning in Large Language Models through Graph-based Synthetic Data
Jiaming Zhou
Abbas Ghaddar
Ge Zhang
Liheng Ma
Yaochen Hu
Soumyasundar Pal
Mark J. Coates
Jianye HAO
B. Wang
Yingxue Zhang
2025-07-23
colmweb.org/COLM/2025/Workshop/XLLM-Reason-Plan (publié)
doi.org
openreview.net
Multiscale Neural PDE Surrogates for Prediction and Downscaling: Application to Ocean Currents
Abdessamad El-Kabid
Loubna Benabbou
Redouane Lguensat
Alex Hernández-García
Accurate modeling of physical systems governed by partial differential equations is a central challenge in scientific computing. In oceanogr… (voir plus)aphy, high-resolution current data are critical for coastal management, environmental monitoring, and maritime safety. However, available satellite products, such as Copernicus data for sea water velocity at ~0.08 degrees spatial resolution and global ocean models, often lack the spatial granularity required for detailed local analyses. In this work, we (a) introduce a supervised deep learning framework based on neural operators for solving PDEs and providing arbitrary resolution solutions, and (b) propose downscaling models with an application to Copernicus ocean current data. Additionally, our method can model surrogate PDEs and predict solutions at arbitrary resolution, regardless of the input resolution. We evaluated our model on real-world Copernicus ocean current data and synthetic Navier-Stokes simulation datasets.
2025-07-23
ArXiv (prépublication)
doi.org
arxiv.org
AURA: A Multi-Modal Medical Agent for Understanding, Reasoning&Annotation
Nima Fathi
Amar Kumar
Tal Arbel
Recent advancements in Large Language Models (LLMs) have catalyzed a paradigm shift from static prediction systems to agentic AI agents capa… (voir plus)ble of reasoning, interacting with tools, and adapting to complex tasks. While LLM-based agentic systems have shown promise across many domains, their application to medical imaging remains in its infancy. In this work, we introduce AURA, the first visual linguistic explainability agent designed specifically for comprehensive analysis, explanation, and evaluation of medical images. By enabling dynamic interactions, contextual explanations, and hypothesis testing, AURA represents a significant advancement toward more transparent, adaptable, and clinically aligned AI systems. We highlight the promise of agentic AI in transforming medical image analysis from static predictions to interactive decision support. Leveraging Qwen-32B, an LLM-based architecture, AURA integrates a modular toolbox comprising: (i) a segmentation suite with phase grounding, pathology segmentation, and anatomy segmentation to localize clinically meaningful regions; (ii) a counterfactual image-generation module that supports reasoning through image-level explanations; and (iii) a set of evaluation tools including pixel-wise difference-map analysis, classification, and advanced state-of-the-art components to assess diagnostic relevance and visual interpretability.
2025-07-21
ArXiv (prépublication)
doi.org
arxiv.org
Enhancing Changepoint Detection: Penalty Learning through Deep Learning Techniques
Tung L. Nguyen
Toby Dylan Hocking
2025-07-21
Statistics and Computing (publié)
doi.org
arxiv.org
Tracing Optimization for Performance Modeling and Regression Detection
Kaveh Shahedi
Heng Li
Maxime Lamothe
Foutse Khomh
Software performance modeling plays a crucial role in developing and maintaining software systems. A performance model analytically describe… (voir plus)s the relationship between the performance of a system and its runtime activities. This process typically examines various aspects of a system's runtime behavior, such as the execution frequency of functions or methods, to forecast performance metrics like program execution time. By using performance models, developers can predict expected performance and thereby effectively identify and address unexpected performance regressions when actual performance deviates from the model's predictions. One common and precise method for capturing performance behavior is software tracing, which involves instrumenting the execution of a program, either at the kernel level (e.g., system calls) or application level (e.g., function calls). However, due to the nature of tracing, it can be highly resource-intensive, making it impractical for production environments where resources are limited. In this work, we propose statistical approaches to reduce tracing overhead by identifying and excluding performance-insensitive code regions, particularly application-level functions, from tracing while still building accurate performance models that can capture performance degradations. By selecting an optimal set of functions to be traced, we can construct optimized performance models that achieve an R-2 score of up to 99% and, sometimes, outperform full tracing models (models using non-optimized tracing data), while significantly reducing the tracing overhead by more than 80% in most cases. Our optimized performance models can also capture performance regressions in our studied programs effectively, demonstrating their usefulness in real-world scenarios. Our approach is fully automated, making it ready to be used in production environments with minimal human effort.
2025-07-21
ACM Transactions on Software Engineering and Methodology (publié)
doi.org
arxiv.org
Corrigendum to "Child- and Proxy-reported Differences in Patient-reported Outcome and Experience Measures in Pediatric Surgery: Systematic Review and Meta-analysis" [Journal of Pediatric Surgery 60 (2025) 162172].
Zanib Nafees
Siena O'Neill
Alexandra Dimmer
Elena Guadagno
Julia Ferreira
Nancy Mayo
Dan Poenaru
2025-07-20
Journal of Pediatric Surgery (publié)
doi.org
Corrigendum to "Virtual Reality for Pediatric Trauma Education - A Preliminary Face and Content Validation Study" [Journal of Pediatric Surgery 60 (2025) 161951].
F. Botelho
Said Ashkar
Shreenik Kundu
TJ Matthews
Elena Guadagno
Dan Poenaru
2025-07-20
Journal of Pediatric Surgery (publié)
doi.org
The impact of statistical adjustment for assay performance on inferences from SARS-CoV-2 serological surveillance studies
Jiacheng Chen
Yuan Yu
Sheila F O’Brien
Carmen L Charlton
Steven J Drews
Jane M Heffernan
Amber M Smith
Yu Nakagama
Yasutoshi Kido
David L Buckeridge
W Alton Russell
Abstract Choice of immunoassay influences population seroprevalence estimates. Post hoc adjustments for assay performance could improve comp… (voir plus)arability of estimates across studies and enable pooled analyses. We assessed post hoc adjustment methods using data from 2021 to 2023 SARS-CoV-2 serosurveillance studies in Alberta, Canada: one that tested 124 008 blood donations using Roche immunoassays (SARS-CoV-2 nucleocapsid total antibody and anti–SARS-CoV-2 S) and another that tested 214 780 patient samples using Abbott immunoassays (SARS-CoV-2 IgG and anti–SARS-CoV-2 S). Comparing datasets, seropositivity for antibodies against nucleocapsid (anti-N) diverged after May 2022 due to differential loss of sensitivity as a function of time since infection. The commonly used Rogan-Gladen adjustment did not reduce this divergence. Regression-based adjustments using the assays’ semiquantitative results produced more similar estimates of anti-N seroprevalence and rolling incidence proportion (proportion of individuals infected in recent months). Seropositivity for antibodies targeting SARS-CoV-2 spike protein was similar without adjustment, and concordance was not improved when applying an alternative, functional threshold. These findings suggest that assay performance substantially impacted population inferences from SARS-CoV-2 serosurveillance studies in the Omicron period. Unlike methods that ignore time-varying assay sensitivity, regression-based methods using the semiquantitative assay resulted in increased concordance in estimated anti-N seropositivity and rolling incidence between cohorts using different assays.
2025-07-18
American Journal of Epidemiology (publié)
doi.org
Bering: joint cell segmentation and annotation for spatial transcriptomics with transferred graph embeddings
Kang Jin
Zuobai Zhang
Ke Zhang
Francesca Viggiani
Claire Callahan
Jian Tang
Bruce J. Aronow
Jian Shu
Single-cell spatial transcriptomics such as in-situ hybridization or sequencing technologies can provide subcellular resolution that enables… (voir plus) the identification of individual cell identities, locations, and a deep understanding of subcellular mechanisms. However, accurate segmentation and annotation that allows individual cell boundaries to be determined remains a major challenge that limits all the above and downstream insights. Current machine learning methods heavily rely on nuclei or cell body staining, resulting in the significant loss of both transcriptome depth and the limited ability to learn latent representations of spatial colocalization relationships. Here, we propose Bering, a graph deep learning model that leverages transcript colocalization relationships for joint noise-aware cell segmentation and molecular annotation in 2D and 3D spatial transcriptomics data. Graph embeddings for the cell annotation are transferred as a component of multi-modal input for cell segmentation, which is employed to enrich gene relationships throughout the process. To evaluate performance, we benchmarked Bering with state-of-the-art methods and observed significant improvement in cell segmentation accuracies and numbers of detected transcripts across various spatial technologies and tissues. To streamline segmentation processes, we constructed expansive pre-trained models, which yield high segmentation accuracy in new data through transfer learning and self-distillation, demonstrating the generalizability of Bering.
2025-07-17
Nature Communications (publié)
doi.org