Portrait of Paul Bertin

Paul Bertin

PhD - Université de Montréal
Supervisor
Research Topics
Causality
Computational Biology
Deep Learning

Publications

A scalable gene network model of regulatory dynamics in single cells
Joseph D Viviano
Alejandro Tejada-Lapuerta
Weixu Wang
Fabian J. Theis
A scalable gene network model of regulatory dynamics in single cells
Joseph D Viviano
Alejandro Tejada-Lapuerta
Weixu Wang
Fabian J. Theis
Causal machine learning for single-cell genomics
Alejandro Tejada-Lapuerta
Hananeh Aliee
Fabian J. Theis
RECOVER identifies synergistic drug combinations in vitro through sequential model optimization
Deepak Sharma
Thomas Gaudelet
Andrew Anighoro
Torsten Gross
Francisco Martínez-Peña
Eileen L. Tang
M.S. Suraj
Cristian Regep
Jeremy B.R. Hayter
Maksym Korablyov
Nicholas Valiante
Mike Tyers
Charles E.S. Roberts
Michael M. Bronstein
Luke L. Lairson
Jake P. Taylor-King
DEUP: Direct Epistemic Uncertainty Prediction
Moksh J. Jain
Hadi Nekoei
Victor I Butoi
Maksym Korablyov
Epistemic Uncertainty is a measure of the lack of knowledge of a learner which diminishes with more evidence. While existing work focuses on… (see more) using the variance of the Bayesian posterior due to parameter uncertainty as a measure of epistemic uncertainty, we argue that this does not capture the part of lack of knowledge induced by model misspecification. We discuss how the excess risk, which is the gap between the generalization error of a predictor and the Bayes predictor, is a sound measure of epistemic uncertainty which captures the effect of model misspecification. We thus propose a principled framework for directly estimating the excess risk by learning a secondary predictor for the generalization error and subtracting an estimate of aleatoric uncertainty, i.e., intrinsic unpredictability. We discuss the merits of this novel measure of epistemic uncertainty, and highlight how it differs from variance-based measures of epistemic uncertainty and addresses its major pitfall. Our framework, Direct Epistemic Uncertainty Prediction (DEUP) is particularly interesting in interactive learning environments, where the learner is allowed to acquire novel examples in each round. Through a wide set of experiments, we illustrate how existing methods in sequential model optimization can be improved with epistemic uncertainty estimates from DEUP, and how DEUP can be used to drive exploration in reinforcement learning. We also evaluate the quality of uncertainty estimates from DEUP for probabilistic image classification and predicting synergies of drug combinations.
RECOVER: sequential model optimization platform for combination drug repurposing identifies novel synergistic compounds in vitro
Deepak Sharma
Thomas Gaudelet
Andrew Anighoro
Torsten Gross
Francisco Martínez-Peña
Eileen L. Tang
S. SurajM
Cristian Regep
Jeremy B.R. Hayter
Maksym Korablyov
N. Valiante
Almer M. van der Sloot
Mike Tyers
Charles E.S. Roberts
Michael M. Bronstein
Luke Lee Lairson
Jake P. Taylor-King