Publications

Leveraging World Model Disentanglement in Value-Based Multi-Agent Reinforcement Learning
Zhizun Wang
David Meger
In this paper, we propose a novel model-based multi-agent reinforcement learning approach named Value Decomposition Framework with Disentang… (voir plus)led World Model to address the challenge of achieving a common goal of multiple agents interacting in the same environment with reduced sample complexity. Due to scalability and non-stationarity problems posed by multi-agent systems, model-free methods rely on a considerable number of samples for training. In contrast, we use a modularized world model, composed of action-conditioned, action-free, and static branches, to unravel the environment dynamics and produce imagined outcomes based on past experience, without sampling directly from the real environment. We employ variational auto-encoders and variational graph auto-encoders to learn the latent representations for the world model, which is merged with a value-based framework to predict the joint action-value function and optimize the overall training objective. We present experimental results in Easy, Hard, and Super-Hard StarCraft II micro-management challenges to demonstrate that our method achieves high sample efficiency and exhibits superior performance in defeating the enemy armies compared to other baselines.
2023-09-07
ArXiv (prépublication)
doi.org
arxiv.org
Alignment of auditory artificial networks with massive individual fMRI brain data leads to generalisable improvements in brain encoding and downstream tasks
Maelle Freteault
Loic Tetrel
Maximilien Le Clei
Lune P Bellec
Nicolas Farrugia
Artificial neural networks trained in the field of artificial intelligence (AI) have emerged as key tools to model brain processes, sparking… (voir plus) the idea of aligning network representations with brain dynamics to enhance performance on AI tasks. While this concept has gained support in the visual domain, we investigate here the feasibility of creating auditory artificial neural models directly aligned with individual brain activity. This objective raises major computational challenges, as models have to be trained directly with brain data, which is typically collected at a much smaller scale than data used to train AI models. We aimed to answer two key questions: (1) Can brain alignment of auditory models lead to improved brain encoding for novel, previously unseen stimuli? (2) Can brain alignment lead to generalisable representations of auditory signals that are useful for solving a variety of complex auditory tasks? To answer these questions, we relied on two massive datasets: a deep phenotyping dataset from the Courtois neuronal modelling project, where six subjects watched four seasons (36 hours) of the Friends TV series in functional magnetic resonance imaging and the HEAR benchmark, a large battery of downstream auditory tasks. We fine-tuned SoundNet, a small pretrained convolutional neural network with ∼2.5M parameters. Aligning SoundNet with brain data from three seasons of Friends led to substantial improvement in brain encoding in the fourth season, extending beyond auditory and visual cortices. We also observed consistent performance gains on the HEAR benchmark, particularly for tasks with limited training data, where brain-aligned models performed comparably to the best-performing models regardless of size. We finally compared individual and group models, finding that individual models often matched or outperformed group models in both brain encoding and downstream task performance, highlighting the data efficiency of fine-tuning with individual brain data. Our results demonstrate the feasibility of aligning artificial neural network representations with individual brain activity during auditory processing, and suggest that this alignment is particularly beneficial for tasks with limited training data. Future research is needed to establish whether larger models can achieve even better performance and whether the observed gains extend to other tasks, particularly in the context of few shot learning.
2023-09-05
bioRxiv (prépublication)
doi.org
Bridging the Gap Between Target Networks and Functional Regularization
Alexandre Piché
Valentin Thomas
Joseph Marino
Gian Maria Marconi
Rafael Pardinas
Christopher Pal
Mohammad Emtiyaz Khan
2023-09-05
TMLR (accepté)
doi.org
openreview.net
Cardiomyocyte orientation recovery at micrometer scale reveals long-axis fiber continuum in heart walls
Drisya Dileep
Tabish A Syed
Tyler FW Sloan
Perundurai S Dhandapany
Kaleem Siddiqi
Minhajuddin Sirajuddin
Coordinated cardiomyocyte contraction drives the mammalian heart to beat and circulate blood. No consensus model of cardiomyocyte geometrica… (voir plus)l arrangement exists, due to the limited spatial resolution of whole heart imaging methods and the piecemeal nature of studies based on histological sections. By combining microscopy and computer vision, we produced the first‐ever three‐dimensional cardiomyocyte orientation reconstruction across mouse ventricular walls at the micrometer scale, representing a gain of three orders of magnitude in spatial resolution. We recovered a cardiomyocyte arrangement aligned to the long‐axis direction of the outer ventricular walls. This cellular network lies in a thin shell and forms a continuum with longitudinally arranged cardiomyocytes in the inner walls, with a complex geometry at the apex. Our reconstruction methods can be applied at fine spatial scales to further understanding of heart wall electrical function and mechanics, and set the stage for the study of micron‐scale fiber remodeling in heart disease.
2023-09-05
The EMBO Journal (publié)
doi.org
Using Multiple Vector Channels Improves E(n)-Equivariant Graph Neural Networks
Daniel Levy
Sékou-Oumar Kaba
Carmelo Gonzales
Santiago Miret
Siamak Ravanbakhsh
2023-09-05
ArXiv (prépublication)
doi.org
arxiv.org
Breaking Barriers to Creative Expression: Co-Designing and Implementing an Accessible Text-to-Image Interface
Atieh Taheri
Mohammad Izadi
Gururaj Shriram
Negar Rostamzadeh
Shaun Kane
Text-to-image generation models have grown in popularity due to their ability to produce high-quality images from a text prompt. One use for… (voir plus) this technology is to enable the creation of more accessible art creation software. In this paper, we document the development of an alternative user interface that reduces the typing effort needed to enter image prompts by providing suggestions from a large language model, developed through iterative design and testing within the project team. The results of this testing demonstrate how generative text models can support the accessibility of text-to-image models, enabling users with a range of abilities to create visual art.
2023-09-04
ArXiv (prépublication)
doi.org
arxiv.org
Deep reinforcement learning for option pricing and hedging under dynamic expectile risk measures
Option Pricing
Saeed Marzban
Erick Delage
Jonathan Yu-Meng Li
2023-09-04
Quantitative Finance (publié)
doi.org
Tidying Up the Conversational Recommender Systems' Biases
Armin Moradi
Golnoosh Farnadi
The growing popularity of language models has sparked interest in conversational recommender systems (CRS) within both industry and research… (voir plus) circles. However, concerns regarding biases in these systems have emerged. While individual components of CRS have been subject to bias studies, a literature gap remains in understanding specific biases unique to CRS and how these biases may be amplified or reduced when integrated into complex CRS models. In this paper, we provide a concise review of biases in CRS by surveying recent literature. We examine the presence of biases throughout the system's pipeline and consider the challenges that arise from combining multiple models. Our study investigates biases in classic recommender systems and their relevance to CRS. Moreover, we address specific biases in CRS, considering variations with and without natural language understanding capabilities, along with biases related to dialogue systems and language models. Through our findings, we highlight the necessity of adopting a holistic perspective when dealing with biases in complex CRS models.
2023-09-04
ArXiv (prépublication)
doi.org
arxiv.org
xSA: A Binary Cross-Entropy Simulated Annealing Polar Decoder
Ryan Seah
Huayi Zhou
Marwan Jalaleddine
Warren J. Gross
Polar decoders such as successive-cancellation and successive-cancellation list decoders are limited by their sequential nature, which leads… (voir plus) to a linear increase in latency with the codeword length. Heuristic based decoders such as quantum annealing have been proposed to overcome this limitation. However, these decoders have shown poor performance when decoding polar codes with more than eight bits. In this paper, we developed new meta-heuristic based polar decoder, called xSA, which uses a new receiver constraint modeled by the binary cross-entropy function. We also propose a method to determine the weights used in a quadratic unconstrained binary optimization (QUBO) function. The polar code is assumed to have been sent across an AWGN channel and we conducted our experiments and simulations using PyQUBO and dwave-neal. Our results show that xSA is able to decode codes of length 16 and 32 with a near-ML FER performance, presenting itself as a promising alternative to traditional polar decoders for real world applications and next generation cellular communications.
2023-09-03
International Symposium on Turbo Codes and Iterative Information Processing (publié)
doi.org
Automated liver segmentation and steatosis grading using deep learning on B-mode ultrasound images
Pedro Vianna
Merve Kulbay
Pamela Boustros
Sara-Ivana Calce
Cassandra Larocque-Rigney
Laurent Patry-Beaudoin
Yi Hui Luo
Muawiz Chaudary
Samuel Kadoury
Bich Nguyen
Emmanuel Montagnon
Eugene Belilovsky
Guy Wolf
Michael Chassé
An Tang
Guy Cloutier
Early detection of nonalcoholic fatty liver disease (NAFLD) is crucial to avoid further complications. Ultrasound is often used for screenin… (voir plus)g and monitoring of hepatic steatosis, however it is limited by the subjective interpretation of images. Computer assisted diagnosis could aid radiologists to achieve objective grading, and artificial intelligence approaches have been tested across various medical applications. In this study, we evaluated the performance of a two-stage hepatic steatosis detection deep learning framework, with a first step of liver segmentation and a subsequent step of hepatic steatosis classification. We evaluated the models on internal and external datasets, aiming to understand the generalizability of the framework. In the external dataset, our segmentation model achieved a Dice score of 0.92 (95% CI: 0.78, 1.00), and our classification model achieved an area under the receiver operating characteristic curve of 0.84 (95% CI: 0.79, 0.89). Our findings highlight the potential benefits of applying artificial intelligence models in NAFLD assessment.
2023-09-02
IUS (publié)
doi.org
122 Is Selenium-75 a Feasible HDR Brachytherapy Source?
Jake Reid
S. Enger
Jonathan Kalinowski
A. Armstrong
John Munro
2023-08-31
Radiotherapy and Oncology (publié)
doi.org
124 Development of a Novel Dosimetry Software for Patient-Specific Intravascular Brachytherapy Treatment Planning on Optical Coherence Tomography Images
Maryam Rahbaran
Jonathan Kalinowski
James Man Git Tsui
Joseph DeCunha
Kevin Croce
Brian Bergmark
Philip Devlin
S. Enger
2023-08-31
Radiotherapy and Oncology (publié)
doi.org