Publications

Gradients without Backpropagation
Atilim Güneş Baydin
Barak A. Pearlmutter
Don Syme
Frank N. Wood
Philip Torr
2022-02-16
ArXiv (preprint)
arxiv.org
arxiv.org
Novel informatics approaches to COVID-19 Research: From methods to applications
Huanan Xu
David L Buckeridge
Yi Wang
P. Tarczy-Hornoch
2022-02-15
Journal of Biomedical Informatics (published)
doi.org
Advanced Diffusion MR Imaging for Multiple Sclerosis in the Brain and Spinal Cord
Masaaki Hori
Tomoko Maekawa
Kouhei Kamiya
Akifumi Hagiwara
Masami Goto
Mariko Yoshida Takemura
Shohei Fujita
Christina Andica
Koji Kamagata
Julien Cohen-Adad
Shigeki Aoki
Diffusion tensor imaging (DTI) has been established its usefulness in evaluating normal-appearing white matter (NAWM) and other lesions that… (see more) are difficult to evaluate with routine clinical MRI in the evaluation of the brain and spinal cord lesions in multiple sclerosis (MS), a demyelinating disease. With the recent advances in the software and hardware of MRI systems, increasingly complex and sophisticated MRI and analysis methods, such as q-space imaging, diffusional kurtosis imaging, neurite orientation dispersion and density imaging, white matter tract integrity, and multiple diffusion encoding, referred to as advanced diffusion MRI, have been proposed. These are capable of capturing in vivo microstructural changes in the brain and spinal cord in normal and pathological states in greater detail than DTI. This paper reviews the current status of recent advanced diffusion MRI for assessing MS in vivo as part of an issue celebrating two decades of magnetic resonance in medical sciences (MRMS), an official journal of the Japanese Society of Magnetic Resonance in Medicine.
2022-02-14
Magnetic Resonance in Medical Sciences (published)
doi.org
Halting Time is Predictable for Large Models: A Universality Property and Average-case Analysis
Courtney Paquette
Bart Van Merriënboer
Fabian Pedregosa
Average-case analysis computes the complexity of an algorithm averaged over all possible inputs. Compared to worst-case analysis, it is more… (see more) representative of the typical behavior of an algorithm, but remains largely unexplored in optimization. One difficulty is that the analysis can depend on the probability distribution of the inputs to the model. However, we show that this is not the case for a class of large-scale problems trained with first-order methods including random least squares and one-hidden layer neural networks with random weights. In fact, the halting time exhibits a universality property: it is independent of the probability distribution. With this barrier for average-case analysis removed, we provide the first explicit average-case convergence rates showing a tighter complexity not captured by traditional worst-case analysis. Finally, numerical simulations suggest this universality property holds for a more general class of algorithms and problems.
2022-02-14
Foundations of Computational Mathematics (published)
doi.org
arxiv.org
Geographic concentration of SARS-CoV-2 cases by social determinants of health in metropolitan areas in Canada: a cross-sectional study
Yiqing Xia
Huiting Ma
Gary Moloney
Héctor Alexander Velásquez García
Monica Sirski
Naveed Z. Janjua
David Vickers
Tyler Williamson
Alan Katz
Kristy C.Y. Yiu
Rafal Kustra
David L. Buckeridge
Marc Brisson
Stefan Baral
Sharmistha Mishra
Mathieu Maheu‐Giroux
The feature of geographical concentration of COVID-19 cases was consistent across CMAs, but the pattern by social determinants varied. Geogr… (see more)aphically-prioritized allocation of resources and services should be tailored to the local drivers of inequalities in transmission in response to SARS-CoV-2’s resurgence.
2022-02-13
CMAJ : Canadian Medical Association Journal (published)
doi.org
Erratum to: Rapid simultaneous acquisition of macromolecular tissue volume, susceptibility, and relaxometry maps (Magn Reson Med. 2022;87:781‐790.)
Fang Frank Yu
Susie Yi Huang
Ashwin Kumar
Thomas Witzel
Congyu Liao
Tanguy Duval
Julien Cohen-Adad
Berkin Bilgic
2022-02-09
Magnetic Resonance in Medicine (published)
doi.org
The Brain-Computer Metaphor Debate Is Useless: A Matter of Semantics
Blake A. Richards
Timothy P. Lillicrap
It is commonly assumed that usage of the word “computer” in the brain sciences reflects a metaphor. However, there is no single definiti… (see more)on of the word “computer” in use. In fact, based on the usage of the word “computer” in computer science, a computer is merely some physical machinery that can in theory compute any computable function. According to this definition the brain is literally a computer; there is no metaphor. But, this deviates from how the word “computer” is used in other academic disciplines. According to the definition used outside of computer science, “computers” are human-made devices that engage in sequential processing of inputs to produce outputs. According to this definition, brains are not computers, and arguably, computers serve as a weak metaphor for brains. Thus, we argue that the recurring brain-computer metaphor debate is actually just a semantic disagreement, because brains are either literally computers or clearly not very much like computers at all, depending on one's definitions. We propose that the best path forward is simply to put the debate to rest, and instead, have researchers be clear about which definition they are using in their work. In some circumstances, one can use the definition from computer science and simply ask, what type of computer is the brain? In other circumstances, it is important to use the other definition, and to clarify the ways in which our brains are radically different from the laptops, smartphones, and servers that surround us in modern life.
2022-02-07
Frontiers in Computer Science (published)
doi.org
Minimizing Entropy to Discover Good Solutions to Recurrent Mixed Integer Programs
Charly Robinson La Rocca
Emma Frejinger
Jean-François Cordeau
Current state-of-the-art solvers for mixed-integer programming (MIP) problems are designed to perform well on a wide range of problems. Howe… (see more)ver, for many real-world use cases, problem instances come from a narrow distribution. This has motivated the development of specialized methods that can exploit the information in historical datasets to guide the design of heuristics. Recent works have shown that machine learning (ML) can be integrated with an MIP solver to inject domain knowledge and efficiently close the optimality gap. This hybridization is usually done with deep learning (DL), which requires a large dataset and extensive hyperparameter tuning to perform well. This paper proposes an online heuristic that uses the notion of entropy to efficiently build a model with minimal training data and tuning. We test our method on the locomotive assignment problem (LAP), a recurring real-world problem that is challenging to solve at scale. Experimental results show a speed up of an order of magnitude compared to a general purpose solver (CPLEX) with a relative gap of less than 2%. We also observe that for some instances our method can discover better solutions than CPLEX within the time limit.
2022-02-06
ArXiv (preprint)
arxiv.org
arxiv.org
RECOVER: sequential model optimization platform for combination drug repurposing identifies novel synergistic compounds in vitro
Paul Bertin
Jarrid Rector-Brooks
Deepak Sharma
Thomas Gaudelet
Andrew Anighoro
Torsten Gross
Francisco Martínez-Peña
Eileen L. Tang
Suraj M S
Cristian Regep
Jeremy B. R. Hayter
Maksym Korablyov
Nicholas Valiante
Almer Van Der Sloot
Mike Tyers
Charles Roberts
Michael M. Bronstein
Luke L. Lairson
Jake P. Taylor-King
Yoshua Bengio
2022-02-06
arXiv (preprint)
doi.org
arxiv.org
Bayesian Neural Networks
Tom Charnock
Laurence Perreault-Levasseur
François Lanusse
In recent times, neural networks have become a powerful tool for the analysis of complex and abstract data models. However, their introducti… (see more)on intrinsically increases our uncertainty about which features of the analysis are model-related and which are due to the neural network. This means that predictions by neural networks have biases which cannot be trivially distinguished from being due to the true nature of the creation and observation of data or not. In order to attempt to address such issues we discuss Bayesian neural networks: neural networks where the uncertainty due to the network can be characterised. In particular, we present the Bayesian statistical framework which allows us to categorise uncertainty in terms of the ingrained randomness of observing certain data and the uncertainty from our lack of knowledge about how data can be created and observed. In presenting such techniques we show how errors in prediction by neural networks can be obtained in principle, and provide the two favoured methods for characterising these errors. We will also describe how both of these methods have substantial pitfalls when put into practice, highlighting the need for other statistical techniques to truly be able to do inference when using neural networks.
2022-02-05
Artificial Intelligence for High Energy Physics (published)
doi.org
arxiv.org
Exploration with Multi-Sample Target Values for Distributional Reinforcement Learning
Michael Teng
Michiel van de Panne
Frank N. Wood
Distributional reinforcement learning (RL) aims to learn a value-network that predicts the full distribution of the returns for a given stat… (see more)e, often modeled via a quantile-based critic. This approach has been successfully integrated into common RL methods for continuous control, giving rise to algorithms such as Distributional Soft Actor-Critic (DSAC). In this paper, we introduce multi-sample target values (MTV) for distributional RL, as a principled replacement for single-sample target value estimation, as commonly employed in current practice. The improved distributional estimates further lend themselves to UCB-based exploration. These two ideas are combined to yield our distributional RL algorithm, E2DC (Extra Exploration with Distributional Critics). We evaluate our approach on a range of continuous control tasks and demonstrate state-of-the-art model-free performance on difficult tasks such as Humanoid control. We provide further insight into the method via visualization and analysis of the learned distributions and their evolution during training.
2022-02-05
ArXiv (preprint)
arxiv.org
arxiv.org
TIML: Task-Informed Meta-Learning for Agriculture
Gabriel Tseng
Hannah Kerner
David Rolnick
Labeled datasets for agriculture are extremely spatially imbalanced. When developing algorithms for data-sparse regions, a natural approach … (see more)is to use transfer learning from data-rich regions. While standard transfer learning approaches typically leverage only direct inputs and outputs, geospatial imagery and agricultural data are rich in metadata that can inform transfer learning algorithms, such as the spatial coordinates of data-points or the class of task being learned. We build on previous work exploring the use of meta-learning for agricultural contexts in data-sparse regions and introduce task-informed meta-learning (TIML), an augmentation to model-agnostic meta-learning which takes advantage of task-specific metadata. We apply TIML to crop type classification and yield estimation, and find that TIML significantly improves performance compared to a range of benchmarks in both contexts, across a diversity of model architectures. While we focus on tasks from agriculture, TIML could offer benefits to any meta-learning setup with task-specific metadata, such as classification of geo-tagged images and species distribution modelling.
2022-02-03
ArXiv (preprint)
arxiv.org
arxiv.org