Publications

Simplicial Embeddings Improve Sample Efficiency in Actor-Critic Agents
Johan Samir Obando Ceron
walter Mayor
Samuel Lavoie
Scott Fujimoto
Aaron Courville
Pablo Samuel Castro
Recent works have proposed accelerating the wall-clock training time of actor-critic methods via the use of large-scale environment parallel… (voir plus)ization; unfortunately, these can sometimes still require large number of environment interactions to achieve a desired level of performance. Noting that well-structured representations can improve the generalization and sample efficiency of deep reinforcement learning (RL) agents, we propose the use of simplicial embeddings: lightweight representation layers that constrain embeddings to simplicial structures. This geometric inductive bias results in sparse and discrete features that stabilize critic bootstrapping and strengthen policy gradients. When applied to FastTD3, FastSAC, and PPO, simplicial embeddings consistently improve sample efficiency and final performance across a variety of continuous- and discrete-control environments, without any loss in runtime speed.
2025-10-15
ArXiv (prépublication)
arxiv.org
arxiv.org
Simplicial Embeddings Improve Sample Efficiency in Actor-Critic Agents
Johan Samir Obando Ceron
walter Mayor
Samuel Lavoie
Scott Fujimoto
Aaron Courville
Pablo Samuel Castro
Recent works have proposed accelerating the wall-clock training time of actor-critic methods via the use of large-scale environment parallel… (voir plus)ization; unfortunately, these can sometimes still require large number of environment interactions to achieve a desired level of performance. Noting that well-structured representations can improve the generalization and sample efficiency of deep reinforcement learning (RL) agents, we propose the use of simplicial embeddings: lightweight representation layers that constrain embeddings to simplicial structures. This geometric inductive bias results in sparse and discrete features that stabilize critic bootstrapping and strengthen policy gradients. When applied to FastTD3, FastSAC, and PPO, simplicial embeddings consistently improve sample efficiency and final performance across a variety of continuous- and discrete-control environments, without any loss in runtime speed.
2025-10-15
ArXiv (prépublication)
arxiv.org
arxiv.org
The Interpolation Constraint in the RV Analysis of M Dwarfs Using Empirical Templates
Dhvani Doshi
Nicolas B. Cowan
Étienne Artigau
René Doyon
André M. Silva
Khaled Al Moulla
Yashar Hezaveh
2025-10-15
The Astronomical Journal (publié)
doi.org
The Interpolation Constraint in the RV Analysis of M-Dwarfs Using Empirical Templates
Dhvani Doshi
Nicolas B. Cowan
E. Artigau
René Doyon
André M. Silva
Khaled Al Moulla
Yashar Hezaveh
Precise radial velocity (pRV) measurements of M-dwarfs in the near-infrared (NIR) rely on empirical templates due to the lack of accurate st… (voir plus)ellar spectral models in this regime. Templates are assumed to approximate the true spectrum when constructed from many observations or in the high signal-to-noise limit. We develop a numerical simulation that generates SPIRou-like pRV observations from PHOENIX spectra, constructs empirical templates, and estimates radial velocities. This simulation solely considers photon noise and evaluates when empirical templates remain reliable for pRV analysis. Our results reveal a previously unrecognized noise source in templates, establishing a fundamental floor for template-based pRV measurements. We find that templates inherently include distortions in stellar line shapes due to imperfect interpolation at the detector's sampling resolution. The magnitude of this interpolation error depends on sampling resolution and RV content. Consequently, while stars with a higher RV content, such as cooler M-dwarfs are expected to yield lower RV uncertainties, their dense spectral features can amplify interpolation errors, potentially biasing RV estimates. For a typical M4V star, SPIRou's spectral and sampling resolution imposes an RV uncertainty floor of 0.5-0.8 m/s, independent of the star's magnitude or the telescope's aperture. These findings reveal a limitation of template-based pRV methods, underscoring the need for improved spectral modeling and better-than-Nyquist detector sampling to reach the next level of RV precision.
2025-10-15
The Astronomical Journal (publié)
doi.org
arxiv.org
Personalized Federated Fine-Tuning of Vision Foundation Models for Healthcare
Adam Tupper
Christian Gagné
2025-10-14
ArXiv (prépublication)
arxiv.org
arxiv.org
Personalized Federated Fine-Tuning of Vision Foundation Models for Healthcare
Adam Tupper
Christian Gagné
Foundation models open up new possibilities for the use of AI in healthcare. However, even when pre-trained on health data, they still need … (voir plus)to be fine-tuned for specific downstream tasks. Furthermore, although foundation models reduce the amount of training data required to achieve good performance, obtaining sufficient data is still a challenge. This is due, in part, to restrictions on sharing and aggregating data from different sources to protect patients'privacy. One possible solution to this is to fine-tune foundation models via federated learning across multiple participating clients (i.e., hospitals, clinics, etc.). In this work, we propose a new personalized federated fine-tuning method that learns orthogonal LoRA adapters to disentangle general and client-specific knowledge, enabling each client to fully exploit both their own data and the data of others. Our preliminary results on real-world federated medical imaging tasks demonstrate that our approach is competitive against current federated fine-tuning methods.
2025-10-14
ArXiv (prépublication)
arxiv.org
arxiv.org
TimelyGPT: Extrapolatable Transformer Pre-training for Long-term Time-Series Forecasting in Healthcare
Ziyang Song
Qincheng Lu
Hao Xu
Ziqi Yang
Mike He Zhu
David Buckeridge
Yue Li
Motivation: Large-scale pre-trained models (PTMs) such as BERT and GPT have recently achieved great success in Natural Language Processing a… (voir plus)nd Computer Vision domains. However, the development of PTMs on healthcare time-series data is lagging behind. This underscores the limitations of the existing transformer-based architectures, particularly their scalability to handle large-scale time series and ability to capture long-term temporal dependencies. Methods: In this study, we present Timely Generative Pre-trained Transformer (TimelyGPT). TimelyGPT employs an extrapolatable position (xPos) embedding to encode trend and periodic patterns into time-series representations. It also integrates recurrent attention and temporal convolution modules to effectively capture global-local temporal dependencies. Materials: We evaluated TimelyGPT on two large-scale healthcare time series datasets corresponding to continuous biosignals and irregularly-sampled time series, respectively: (1) the Sleep EDF dataset consisting of over 1.2 billion timesteps; (2) the longitudinal healthcare administrative database PopHR, comprising 489,000 patients randomly sampled from the Montreal population. Results: In forecasting continuous biosignals, TimelyGPT achieves accurate extrapolation up to 6,000 timesteps of body temperature during the sleep stage transition, given a short look-up window (i.e., prompt) containing only 2,000 timesteps. For irregularly-sampled time series, TimelyGPT with a proposed time-specific inference demonstrates high top recall scores in predicting future diagnoses using early diagnostic records, effectively handling irregular intervals between clinical records. Together, we envision TimelyGPT to be useful in various health domains, including long-term patient health state forecasting and patient risk trajectory prediction. Availability: The open-sourced code is available at Github.
2025-10-14
Health Information Science and Systems (publié)
doi.org
arxiv.org
Timelygpt: extrapolatable transformer pre-training for long-term time-series forecasting in healthcare
Ziyang Song
Qincheng Lu
Hao Xu
Ziqi Yang
Mike He Zhu
David Buckeridge
Yue Li
2025-10-14
Health Information Science and Systems (publié)
www.ncbi.nlm.nih.gov
Disease-Specific Prediction of Missense Variant Pathogenicity with DNA Language Models and Graph Neural Networks
Mohamed Ghadie
Sameer Sardaar
Yannis Trakadis
2025-10-13
Bioengineering (publié)
doi.org
Disease-Specific Prediction of Missense Variant Pathogenicity with DNA Language Models and Graph Neural Networks
Mohamed Ghadie
Sameer Sardaar
Yannis Trakadis
2025-10-13
Bioengineering (publié)
doi.org
Disease-Specific Prediction of Missense Variant Pathogenicity with DNA Language Models and Graph Neural Networks
Mohamed Ghadie
Sameer Sardaar
Yannis Trakadis
Accurate prediction of the impact of genetic variants on human health is of paramount importance to clinical genetics and precision medicine… (voir plus). Recent machine learning (ML) studies have tried to predict variant pathogenicity with different levels of success. However, most missense variants identified on a clinical basis are still classified as variants of uncertain significance (VUS). Our approach allows for the interpretation of a variant for a specific disease and, thus, for the integration of disease-specific domain knowledge. We utilize a comprehensive knowledge graph, with 11 types of interconnected biomedical entities at diverse biomolecular and clinical levels, to classify missense variants from ClinVar. We use BioBERT to generate embeddings of biomedical features for each node in the graph, as well as DNA language models to embed variant features directly from genomic sequence. Next, we train a two-stage architecture consisting of a graph convolutional neural network to encode biological relationships. A neural network is then used as the classifier to predict disease-specific pathogenicity of variants, essentially predicting edges between variant and disease nodes. We compare performance across different versions of our model, obtaining prediction-balanced accuracies as high as 85.6% (sensitivity: 90.5%; NPV: 89.8%) and discuss how our work can inform future studies in this area.
2025-10-13
Bioengineering (publié)
doi.org
Iterative Amortized Inference: Unifying In-Context Learning and Learned Optimizers
Sarthak Mittal
Divyat Mahajan
Guillaume Lajoie
Mohammad Pezeshki
Modern learning systems increasingly rely on amortized learning - the idea of reusing computation or inductive biases shared across tasks to… (voir plus) enable rapid generalization to novel problems. This principle spans a range of approaches, including meta-learning, in-context learning, prompt tuning, learned optimizers and more. While motivated by similar goals, these approaches differ in how they encode and leverage task-specific information, often provided as in-context examples. In this work, we propose a unified framework which describes how such methods differ primarily in the aspects of learning they amortize - such as initializations, learned updates, or predictive mappings - and how they incorporate task data at inference. We introduce a taxonomy that categorizes amortized models into parametric, implicit, and explicit regimes, based on whether task adaptation is externalized, internalized, or jointly modeled. Building on this view, we identify a key limitation in current approaches: most methods struggle to scale to large datasets because their capacity to process task data at inference (e.g., context length) is often limited. To address this, we propose iterative amortized inference, a class of models that refine solutions step-by-step over mini-batches, drawing inspiration from stochastic optimization. Our formulation bridges optimization-based meta-learning with forward-pass amortization in models like LLMs, offering a scalable and extensible foundation for general-purpose task adaptation.
2025-10-13
ArXiv (prépublication)
arxiv.org
arxiv.org