Publications

Designing Biological Sequences via Meta-Reinforcement Learning and Bayesian Optimization
Leo Feng
Padideh Nouri
Aneri Muni
Yoshua Bengio
Pierre-Luc Bacon
The ability to accelerate the design of biological sequences can have a substantial impact on the progress of the medical field. The problem… (see more) can be framed as a global optimization problem where the objective is an expensive black-box function such that we can query large batches restricted with a limitation of a low number of rounds. Bayesian Optimization is a principled method for tackling this problem. However, the astronomically large state space of biological sequences renders brute-force iterating over all possible sequences infeasible. In this paper, we propose MetaRLBO where we train an autoregressive generative model via Meta-Reinforcement Learning to propose promising sequences for selection via Bayesian Optimization. We pose this problem as that of finding an optimal policy over a distribution of MDPs induced by sampling subsets of the data acquired in the previous rounds. Our in-silico experiments show that meta-learning over such ensembles provides robustness against reward misspecification and achieves competitive results compared to existing strong baselines.
2022-09-12
ArXiv (preprint)
doi.org
arxiv.org
Measuring Commonality in Recommendation of Cultural Content: Recommender Systems to Enhance Cultural Citizenship
Andres Ferraro
Gustavo Ferreira
Fernando Diaz
Georgina Born
2022-09-12
Proceedings of the 16th ACM Conference on Recommender Systems (published)
doi.org
arxiv.org
Towards Fair Federated Recommendation Learning: Characterizing the Inter-Dependence of System and Data Heterogeneity
Kiwan Maeng
Haiyu Lu
Luca Melis
John Nguyen
Michael G. Rabbat
Carole-Jean Wu
2022-09-12
Proceedings of the 16th ACM Conference on Recommender Systems (published)
doi.org
arxiv.org
Learning Multi-Objective Curricula for Robotic Policy Learning
Jikun Kang
Miao Liu
Abhinav Gupta
Christopher Pal
Xue Liu
Jie Fu
2022-09-09
Conference on Robot Learning (poster)
doi.org
proceedings.mlr.press
A neurodynamic model of inter-brain coupling in the gamma band
Quentin Moreau
Lena Adel
Caitriona Douglas
Ranjbaran Ghazaleh
Guillaume Dumas
2022-09-06
Journal of Neurophysiology (published)
doi.org
Unifying Generative Models with GFlowNets
Dinghuai Zhang
Ricky T. Q. Chen
Nikolay Malkin
Yoshua Bengio
There are many frameworks for deep generative modeling, each often presented with their own specific training algorithms and inference metho… (see more)ds. Here, we demonstrate the connections between existing deep generative models and the recently introduced GFlowNet framework, a probabilistic inference machine which treats sampling as a decision-making process. This analysis sheds light on their overlapping traits and provides a unifying viewpoint through the lens of learning with Markovian trajectories. Our framework provides a means for unifying training and inference algorithms, and provides a route to shine a unifying light over many generative models. Beyond this, we provide a practical and experimentally verified recipe for improving generative modeling with insights from the GFlowNet perspective.
2022-09-05
ArXiv (preprint)
doi.org
arxiv.org
User Experience of a Computer-Based Decision Aid for Prenatal Trisomy Screening: Mixed Methods Explanatory Study
Titilayo Tatiana Agbadje
Chantale Pilon
Pierre Bérubé
Jean-Claude Forest
François Rousseau
S. A. Rahimi
Yves Giguère
France Légaré
2022-09-05
JMIR Pediatrics and Parenting (published)
doi.org
Re-expression of CA1 and entorhinal activity patterns preserves temporal context memory at long timescales
Futing Zou
Wanjia Guo
Emily J. Allen
Yihan Wu
Ian Charest
Thomas Naselaris
Kendrick Kay
Brice A. Kuhl
J. Benjamin Hutchinson
Sarah DuBrow
Converging, cross-species evidence indicates that memory for time is supported by hippocampal area CA1 and entorhinal cortex. However, limit… (see more)ed evidence characterizes how these regions preserve temporal memories over long timescales (e.g., months). At long timescales, memoranda may be encountered in multiple temporal contexts, potentially creating interference. Here, using 7T fMRI, we measured CA1 and entorhinal activity patterns as human participants viewed thousands of natural scene images distributed, and repeated, across many months. We show that memory for an image’s original temporal context was predicted by the degree to which CA1/entorhinal activity patterns from the first encounter with an image were re-expressed during re-encounters occurring minutes to months later. Critically, temporal memory signals were dissociable from predictors of recognition confidence, which were carried by distinct medial temporal lobe expressions. These findings suggest that CA1 and entorhinal cortex preserve temporal memories across long timescales by coding for and reinstating temporal context information.
2022-09-02
bioRxiv (preprint)
doi.org
Digitalization and the Anthropocene
Felix Creutzig
Daron Acemoglu
Xuemei Bai
Paul N. Edwards
Marie Josefine Hintz
Lynn H. Kaack
Siir Kilkis
Stefanie Kunkel
Amy Luers
Nikola Milojevic-Dupont
Dave Rejeski
Jürgen Renn
David Rolnick
Christoph Rosol
Daniela Russ
Thomas Turnbull
Elena Verdolini
Felix Wagner
Charlie Wilson
Aicha Zekar … (see 1 more)
Marius Zumwald
Great claims have been made about the benefits of dematerialization in a digital service economy. However, digitalization has historically i… (see more)ncreased environmental impacts at local and planetary scales, affecting labor markets, resource use, governance, and power relationships. Here we study the past, present, and future of digitalization through the lens of three interdependent elements of the Anthropocene: ( a) planetary boundaries and stability, ( b) equity within and between countries, and ( c) human agency and governance, mediated via ( i) increasing resource efficiency, ( ii) accelerating consumption and scale effects, ( iii) expanding political and economic control, and ( iv) deteriorating social cohesion. While direct environmental impacts matter, the indirect and systemic effects of digitalization are more profoundly reshaping the relationship between humans, technosphere and planet. We develop three scenarios: planetary instability, green but inhumane, and deliberate for the good. We conclude with identifying leverage points that shift human–digital–Earth interactions toward sustainability.
2022-09-01
Annual Review Environment and Resources (published)
doi.org
Researcher perspectives on ethics considerations in epigenetics: an international survey
Charles Dupras
Terese Knoppers
Nicole Palmour
Elisabeth Beauchamp
Stamatina Liosi
Reiner Siebert
Alison May Berner
Stephan Beck
Ian Charest
Yann Joly
Over the past decade, bioethicists, legal scholars and social scientists have started to investigate the potential implications of epigeneti… (see more)c research and technologies on medicine and society. There is growing literature discussing the most promising opportunities, as well as arising ethical, legal and social issues (ELSI). This paper explores the views of epigenetic researchers about some of these discussions. From January to March 2020, we conducted an online survey of 189 epigenetic researchers working in 31 countries. We questioned them about the scope of their field, opportunities in different areas of specialization, and ELSI in the conduct of research and knowledge translation. We also assessed their level of concern regarding four emerging non-medical applications of epigenetic testing—i.e., in life insurance, forensics, immigration and direct-to-consumer testing. Although there was strong agreement on DNA methylation, histone modifications, 3D structure of chromatin and nucleosomes being integral elements of the field, there was considerable disagreement on transcription factors, RNA interference, RNA splicing and prions. The most prevalent ELSI experienced or witnessed by respondents were in obtaining timely access to epigenetic data in existing databases, and in the communication of epigenetic findings by the media. They expressed high levels of concern regarding non-medical applications of epigenetics, echoing cautionary appraisals in the social sciences and humanities literature.
2022-09-01
Clinical Epigenetics (published)
doi.org
Small, correlated changes in synaptic connectivity may facilitate rapid motor learning
Barbara Feulner
Matthew G Perich
Raeed H. Chowdhury
Lee Miller
Juan A. Gallego
Claudia Clopath
2022-09-01
Nature Communications (published)
doi.org
Abstract 25 ECTmatch: Optimizing Small-Scale Cord Blood Banking Through HLA Analysis
Maude Dumont-Lagacé
Albert Feghaly
Gabrielle Thauvette
Marie-Christine Meunier
William Lemieux
Diane Fournier
Lucie Richard
Sébastien Lemieux
Guy Sauvageau
Sandra Cohen
Abstract Introduction Cord blood (CB) banks have had to rely on large inventories of CB units to try to serve the largest possible proportio… (see more)n of the population, all the while prioritizing collection of non-Caucasian ethnic groups. However, due to the high linkage disequilibrium of HLA genes and the high frequency of several HLA alleles in the population, CB banks contain hundreds of CB units that could be matched to the same patients, making the inventory somewhat redundant from a clinical standpoint. Objective ExCellThera developed ECTmatch, an algorithm dedicated to optimizing the selection of CB units based on in-depth HLA analysis in order to maximize the efficiency of the bank to suitably match the largest proportion of subjects within a small pool of donors. Methods The performance of ECTmatch was evaluated in a simulation aiming to select 100 CB units from the Héma-Québec CB bank that satisfied an arbitrary minimal cell content criteria of 120 × 107 TNC and 6 × 106 CD34+ cells (n = 2,987). Selection was performed to optimize matching for the Quebec population, with a minimal HLA-match of 5/8 for HLA-A, -B, -C, and -DRB1. Results ECTmatch provides a suitably matched donor for 71.5% of the Quebec population, compared with only 45.0% (±2.4%) with random selection. Because patients who require a CB transplant tend to have rarer HLAs, the performance of ECTmatch was evaluated for this specific subset of patients (n = 62). Again, ECTmatch outperformed random selection, by providing a donor for 54.8% of patients, compared with only 29.9% with random selection. Finally, while ECTmatch was developed to optimize CB selection specifically for the Quebec population, it still outperformed random selection for subjects from the other Canadian provinces or the USA. Discussion By selecting CB units based on HLA profiles, ECTmatch allows the creation of a highly useful inventory with a very low number of CB units. This approach to small-scale CB banking can be adapted to different population subsets and could be used to select a subset of CB units for pre-release for immediate clinical availability or for the creation of a pre-expanded CB inventory with maximal population coverage.
2022-08-31
Stem Cells Translational Medicine (published)
doi.org