Publications

Online HD-tRNS over the Right Temporoparietal Junction Enhances Mentalizing during Social Interactions
Vincent Chamberland
Quentin Moreau
Lisane Moses
Gabriela Milanova
Guillaume Dumas
2025-12-31
Neuromodulation: Technology at the Neural Interface (published)
doi.org
Operationalizing the Superficial Alignment Hypothesis via Task Complexity
Tomás Vergara Browne
Darshan Patil
Ivan Titov
Siva Reddy
Tiago Pimentel
Marius Mosbach
The superficial alignment hypothesis (SAH) posits that large language models learn most of their knowledge during pre-training, and that pos… (see more)t-training merely surfaces this knowledge. The SAH, however, lacks a precise definition, which has led to (i) different and seemingly orthogonal arguments supporting it, and (ii) important critiques to it. We propose a new metric called **Task Complexity**: the length of the shortest program that achieves a target performance on a task. In this framework, the SAH claims that pre-trained models drastically reduce the task complexity of achieving high performance on many tasks. Our definition unifies prior arguments supporting the SAH, interpreting them as different strategies to find such short programs. Experimentally, we estimate task complexities of mathematical reasoning, machine translation, and instruction following tasks and show that their respective task complexities can be remarkably low when conditioned on a pre-trained model. Further, we find that pre-training enables access to strong performances on our tasks, but it can require programs of gigabytes of length to access them. Post-training, on the other hand, collapses the complexity of reaching this same performance by several orders of magnitude. Overall, our results highlight that task adaptation can require remarkably little information—often just a few kilobytes.
2025-12-31
International Conference on Machine Learning (Accept (regular))
openreview.net
openreview.net
Optimizing User Profiles via Contextual Bandits for Retrieval-Augmented LLM Personalization
Linfeng Du
Ye Yuan
Zichen Zhao
Fuyuan Lyu
Emiliano Penaloza
Xiuying Chen
Zipeng Sun
Jikun Kang
Laurent Charlin
Xue Liu
Haolun Wu
Large Language Models (LLMs) excel at general-purpose tasks, yet adapting their responses to individual users remains challenging. Retrieval… (see more) augmentation provides a lightweight alternative to fine-tuning by conditioning LLMs on user history records, and existing approaches typically select these records based on semantic relevance. We argue that relevance serves as an unreliable proxy for utility: a record may be semantically similar to a query yet fail to improve generation quality or even degrade it due to redundancy or conflicting information. To bridge this gap, we propose PURPLE, a contextual bandit framework that oPtimizes UseR Profiles for Llm pErsonalization. In contrast to a greedy selection of the most relevant records, PURPLE treats profile construction as a set generation process and utilizes a Plackett-Luce ranking model to capture complex inter-record dependencies. By training with dense feedback provided by the likelihood of the reference response, our method aligns retrieval directly with generation quality. Extensive experiments on nine personalization tasks demonstrate that PURPLE consistently outperforms strong heuristic and retrieval-augmented baselines in both effectiveness and efficiency, establishing a principled and scalable solution for optimizing user profiles.
2025-12-31
arXiv (preprint)
doi.org
arxiv.org
PAC-X: Fuzzy Explainable AI for Multi-Class Malware Detection
Mohd Saqib
Benjamin C. M. Fung
Philippe Charland
2025-12-31
IEEE Trans. Fuzzy Syst. (published)
doi.org
PGT: Procedurally Generated Tasks for improving fine-grained understanding in MLLMs
Rim Assouel
Amir Bar
Michal Drozdzal
Adriana Romero-Soriano
Despite remarkable progress in Multimodal Large Language Models (MLLMs), these models still struggle with fine-grained understanding tasks. … (see more)In this work, we propose **Procedurally Generated Tasks (PGT)** a simple data-driven framework that serves a dual purpose: inducing fine-grained visual understanding and acting as a low-cost diagnostic tool to identify the source of perception failures. By overlaying unambiguous geometric primitives on images, PGT generate additional dense supervision that disentangles visual grounding capability from semantic priors. Extensive experiments on relational, quantitative, and 3D/depth understanding benchmarks show that PGT yields remarkable gains across diverse architectures. Instruction tuning MLLMs on LLaVA-v1.5-Instruct augmented with PGT data results in improvements of up to +20\% on the What’sUp benchmark and +13.3\% on CV-Bench-2D, while maintaining general perception capabilities. Moreover, finetuning state-of-the-art MLLMs on PGT data leads to boosts of up to +5.5\% on What’sUp and +8.3\% on CV-Bench-2D. These findings demonstrate that PGT effectively address the bottleneck of fine-grained perception, revealing that many spatial reasoning deficits stem from inadequate supervision signals rather than inherent architectural or resolution limitations.
2025-12-31
International Conference on Machine Learning (Accept (regular))
openreview.net
openreview.net
PheCode-guided multi-modal topic modeling of electronic health records improves disease incidence prediction and GWAS discovery from UK Biobank
Ziqi Yang
Ziyang Song
Shadi Zabad
Marc-André Legault
Yue Li
Phenome-wide association studies rely on disease definitions derived from diagnostic codes, often failing to leverage the full richness of e… (see more)lectronic health records (EHR). We present MixEHR-SAGE, a PheCode-guided multi-modal topic model that integrates diagnoses, procedures, and medications to enhance phenotyping from large-scale EHRs. By combining expert-informed priors with probabilistic inference, MixEHR-SAGE identifies over 1000 interpretable phenotype topics from UK Biobank data. Applied to 350 000 individuals with high-quality genetic data, MixEHR-SAGE-derived risk scores accurately predict incident type 2 diabetes (T2D) and leukemia diagnoses. Subsequent genome-wide association studies using these continuous risk scores uncovered novel disease-associated loci, including PPP1R15A for T2D and JMJD6/SRSF2 for leukemia, that were missed by traditional binary case definitions. These results highlight the potential of probabilistic phenotyping from multi-modal EHRs to improve genetic discovery. The MixEHR-SAGE software is publicly available at: https://github.com/li-lab-mcgill/MixEHR-SAGE.
2025-12-31
Briefings in Bioinformatics (published)
doi.org
Point Cloud Video Modeling With Progressive Prior Knowledge Guidance and Adaptive Neighboring Aggregation
Jingkun Yan
Hongwei Ge
Mingze Cui
Chunguo Wu
Mingde Zhao
Point cloud video modeling not only has to address the natural irregularity of point clouds, but also the challenge of capturing spatial and… (see more) temporal representation simultaneously. Current methods attempt to approximate the temporal dimension using several 3-D point cloud frame sequences but struggle in sparser conditions. Accurate point trajectory tracking is crucial for effectively capturing temporal dynamics, as point positions across different frames are often inconsistent, especially during rapid motion or at low frame rates. Conventional point tube operations aggregate motion features over fixed time windows but fail to capture rapidly changing scenes. Implicit tracking techniques are limited by quadratic time complexity, which restricts their practical use. In this article, we propose a native 4-D framework (N4DF) that guides the network to learn spatio-temporal dynamics from a native 4-D perspective. Furthermore, we devise a dynamic point spatio-temporal (DPST) convolution to adaptively select the optimal point-tracking strategy, which constructs local plane regions in anchor frames and propagates them to neighboring frames to evaluate point cross-frame movement distances. To further enhance the global modeling power of N4DF, we develop a dynamic self-tracking re-encoding (DSTR) module that employs point-wise self-attention to search for relevant points across the entire video. Compared with the recent 4-D modeling methods, N4DF demonstrates superior performance on MSR-Action3D and NTU RGB+D for action recognition (+0.7% and +1.2% accuracy, respectively), on HOI4D for action segmentation (+1% accuracy), and on Synthia 4-D and nuScenes-lidarseg for semantic segmentation (+0.49% and +1.7% mIoU, respectively). Our N4DF shows greater robustness at low frame-rate settings due to native 4-D modeling and adaptive tracking, making it suitable for tracking fast-moving objects in future real-time scenarios.
2025-12-31
IEEE Transactions on Neural Networks and Learning Systems (published)
doi.org
Position: Agentic AI Systems should be making Bayes-Consistent Decisions
Theodore Papamarkou
Pierre Alquier
Matthias Bauer
Wray Buntine
A. Davison
Gintare Karolina Dziugaite
Maurizio Filippone
Andrew Y. K. Foong
Vincent Fortuin
Dimitris Fouskakis
Eyke Hüllermeier
Theofanis Karaletsos
Mohammad Emtiyaz Khan
Nikita Kotelevskii
S. Lahlou
Yingzhen Li
F Liu
Clare Lyle
Thomas Möllenhoff
Konstantina Palla … (see 9 more)
Maxim Panov
Yusuf Sale
Kajetan Schweighofer
Artem Shelmanov
Siddharth Swaroop
Martin Trapp
Willem Waegeman
Andrew Gordon Wilson
Alexey Zaytsev
2025-12-31
SSRN Electronic Journal (accepted)
doi.org
Position: Causality is Key for Interpretability Claims to Generalise
Shruti Joshi
Aaron Mueller
David Klindt
Wieland Brendel
Dhanya Sridhar
Patrik Reizinger
Interpretability research on large language models (LLMs) has produced methods that align model components to high-level concepts, yet their… (see more) use has been accompanied by recurring failures: findings that do not generalise, and causal language that outruns the evidence. Our position is that Pearl’s causal hierarchy formally defines what constitutes a good alignment, what data or assumptions it requires, and what inferences it supports. Specifically, observations of model behaviour support only associational claims; interventions enable cause-effect claims, but not necessarily predictions of model behaviour; counterfactuals, or predictions of behaviour on unseen examples, are often unverifiable in current studies. We show how interpretability research can benefit from causal representation learning (CRL), which provides tools for provably extracting semantic variables and their relationships from activations, and outline practical requirements for generalisable insights: robustness to distribution shifts, sensitivity to assumptions, and compositionality of interventions. Our diagnostic framework helps practitioners select appropriate methods and mitigate failures to ensure that claims match evidence and findings generalise.
2025-12-31
International Conference on Machine Learning (Accept (regular))
openreview.net
openreview.net
Position: Collusion Risks Among AI Reasoning Agents Justify Certification Requirements for Making Market Decisions
Matthew Riemer
Tommaso Tosato
Maximilian Puelma Touzel
Amin Memarian
Guillaume Dumas
Glen Berseth
Irina Rish
This position paper argues that AI agents with chain-of-thought reasoning capabilities are predisposed to exhibit collusive behavior and sho… (see more)uld be required to obtain behavioral certification before making decisions that affect economic markets. This is because integrating these agents into society could collapse the legal evidentiary distinction between competition and collusion among independent firms without eroding the economic harm distinction. Experiments with DeepSeek-R1 agents in the Bertrand oligopoly pricing domain reveal a tendency towards tacit collusion that persists even when humans prompt the agents not to collude. We further show that the chain-of- thought of these agents can be steered toward either extremely collusive or highly competitive behavior in a way that is not semantically detectable by another LLM analyzing the reasoning traces. As a result, deploying reasoning agents for market decisions leads to collusive economic outcomes without any evidence of conspiracy or intent. Thus, certification based on observed behavior in representative situations is necessary to prevent collusion. We provide preliminary evidence that such agents can be steered in a generalizable way toward efficient competitive equilibria. However, developing a comprehensive behavioral certification will be required before these models can be deployed in real-world markets while ensuring their stability and efficiency.
2025-12-31
International Conference on Machine Learning (Accept (regular))
doi.org
openreview.net
Position: Irresponsible AI: big tech’s influence on AI research and associated impacts
Alex Hernández-García
Alexandra Volokhova
Ezekiel Williams
Dounia Kabakibo
Mélisande Teng
The accelerated development, deployment and adoption of artificial intelligence systems has been fuelled by the increasing presence of big t… (see more)ech in the AI field. This trend has been accompanied by growing ethical concerns and intensified societal and environmental impacts. This position paper argues that irresponsible AI development is strongly driven by big tech's influence and involvement in the field. We develop this argument by laying out the factors through which this influence leads to irresponsible AI. First, we examine the growing and disproportionate influence of big tech in AI research and argue that its drive for scaling and general-purpose systems is fundamentally at odds with the responsible, ethical, and sustainable development of AI. Second, we review key current environmental and societal negative impacts of AI and trace their connections to big tech's influence. Third, we discuss the underlying economic forces driving big tech's actions. Finally, as a call to action, we highlight the need for AI researchers to counter big tech's influence, and review and propose strategies that build on the responsibility of implicated actors and collective action.
2025-12-31
International Conference on Machine Learning (Accept (spotlight))
openreview.net
openreview.net
Position: LLM-Safety Evaluations Lack Robustness
Tim Beyer
Sophie Xhonneux
Simon Geisler
Gauthier Gidel
Leo Schwinn
Stephan Günnemann
In this position paper, we argue that current safety alignment research efforts for large language models are hindered by many intertwined s… (see more)ources of noise, such as small datasets, methodological inconsistencies, and unreliable evaluation setups. This can, at times, make it impossible to evaluate and compare attacks and defenses fairly, thereby slowing research progress. We systematically analyze the LLM safety evaluation pipeline, covering dataset curation, optimization strategies for automated red-teaming, response generation, and response evaluation using LLM judges. At each stage, we identify key issues and highlight their practical impact. We also propose a set of guidelines for reducing noise and bias in evaluations of future attack and defense papers. Lastly, we offer an opposing perspective, highlighting practical reasons for existing limitations. We believe that addressing the outlined problems in future research will improve the field’s ability to generate easily comparable results and make measurable progress.
2025-12-31
International Conference on Machine Learning (Accept (regular))
openreview.net
openreview.net