Doina Precup

Sumana Basu

Collaborateur·rice alumni - McGill

Co-superviseur⋅e :

Adriana Romero Soriano

Collaborateur·rice alumni - McGill

Lynn Cherif

Collaborateur·rice alumni - McGill

Co-superviseur⋅e :

Doctorat - McGill

Co-superviseur⋅e :

Doctorat - McGill

Superviseur⋅e principal⋅e :

David Meger

Jonathan Colaço Carr

Maîtrise recherche - McGill

Superviseur⋅e principal⋅e :

Prakash Panangaden

Élodie Coté-Gauthier

Collaborateur·rice de recherche - McGill

Co-superviseur⋅e :

Isabeau Prémont-Schwarz

Franco Del Balso

Collaborateur·rice de recherche - UdeM

Jesse Farebrother

Doctorat - McGill

Superviseur⋅e principal⋅e :

Marc Gendron-Bellemare

Doctorat - McGill

Superviseur⋅e principal⋅e :

Collaborateur·rice de recherche - Birla Institute of Technology

Howard Huang

Doctorat - McGill

Haque Ishfaq

Collaborateur·rice alumni - McGill

Mohammad Sami Nur Islam Islam

Maîtrise recherche - McGill

Arushi Jain

Collaborateur·rice alumni - McGill

Doctorat - Polytechnique

Doctorat - McGill

Postdoctorat - McGill

Jonathan Lebensold

Collaborateur·rice alumni - McGill

Collaborateur·rice alumni - McGill

Ray Luo

Doctorat - McGill

Superviseur⋅e principal⋅e :

G McCracken

Doctorat - McGill

Nazanin Mohammadi Sepahvand

Collaborateur·rice alumni - McGill

Shahrad Mohammadzadeh

Maîtrise recherche - McGill

Superviseur⋅e principal⋅e :

Gabriela Moisescu-Pareja

Collaborateur·rice de recherche - McGill

Co-superviseur⋅e :

Irina Rish

Padideh Nouri

Doctorat - UdeM

Co-superviseur⋅e :

Doctorat - McGill

Co-superviseur⋅e :

Nate Rahn

Doctorat - McGill

Superviseur⋅e principal⋅e :

Marc Gendron-Bellemare

Manoosh Samiei

Doctorat - McGill

Co-superviseur⋅e :

Doctorat - McGill

Co-superviseur⋅e :

Doctorat - McGill

Nishanth Anand Vemgal

Doctorat - McGill

Doctorat - McGill

Co-superviseur⋅e :

Samira Ebrahimi Kahou

Zihan Wang

Doctorat - McGill

Skipper : combiner l’abstraction spatiale et temporelle afin d’améliorer la généralisation

Guangyuan Wang

Stagiaire de recherche - McGill

Steve Wen

Maîtrise recherche - McGill

Co-superviseur⋅e :

Gregory Dudek

Zijing Wu

Doctorat - McGill

Superviseur⋅e principal⋅e :

Doctorat - McGill

Harry Zhao

Collaborateur·rice alumni - McGill

Co-superviseur⋅e :

Billets de blogue

Generic thumbnail for Mila Blog articles.

22 février 2024

par

Mingde Harry Zhao

Safa Alver

Harm van Seijen

Romain Laroche

Doina Precup

Yoshua Bengio

Lire l'article

Publications

Combined Reinforcement Learning via Abstract Representations

Vincent François-Lavet

Yoshua Bengio

Joelle Pineau

In the quest for efficient and robust reinforcement learning methods, both model-free and model-based approaches offer advantages. In this p… (voir plus)aper we propose a new way of explicitly bridging both approaches via a shared low-dimensional learned encoding of the environment, meant to capture summarizing abstractions. We show that the modularity brought by this approach leads to good generalization while being computationally efficient, with planning happening in a smaller latent state space. In addition, this approach recovers a sufficient low-dimensional representation of the environment, which opens up new strategies for interpretable AI, exploration and transfer learning.

2019-07-16

Proceedings of the AAAI Conference on Artificial Intelligence (publié)

Learning Options with Interest Functions

Khimya Khetarpal

Learning temporal abstractions which are partial solutions to a task and could be reused for solving other tasks is an ingredient that can h… (voir plus)elp agents to plan and learn efficiently. In this work, we tackle this problem in the options framework. We aim to autonomously learn options which are specialized in different state space regions by proposing a notion of interest functions, which generalizes initiation sets from the options framework for function approximation. We build on the option-critic framework to derive policy gradient theorems for interest functions, leading to a new interest-option-critic architecture.

2019-07-16

Proceedings of the AAAI Conference on Artificial Intelligence (publié)

Leveraging Observations in Bandits: Between Risks and Benefits.

Andrei Lupu

Audrey Durand

Imitation learning has been widely used to speed up learning in novice agents, by allowing them to leverage existing data from experts. Allo… (voir plus)wing an agent to be influenced by external observations can benefit to the learning process, but it also puts the agent at risk of following sub-optimal behaviours. In this paper, we study this problem in the context of bandits. More specifically, we consider that an agent (learner) is interacting with a bandit-style decision task, but can also observe a target policy interacting with the same environment. The learner observes only the target’s actions, not the rewards obtained. We introduce a new bandit optimism modifier that uses conditional optimism contingent on the actions of the target in order to guide the agent’s exploration. We analyze the effect of this modification on the well-known Upper Confidence Bound algorithm by proving that it preserves a regret upper-bound of order O(lnT), even in the presence of a very poor target, and we derive the dependency of the expected regret on the general target policy. We provide empirical results showing both great benefits as well as certain limitations inherent to observational learning in the multi-armed bandit setting. Experiments are conducted using targets satisfying theoretical assumptions with high probability, thus narrowing the gap between theory and application.

2019-07-16

Proceedings of the AAAI Conference on Artificial Intelligence (publié)

Self-supervised Learning of Distance Functions for Goal-Conditioned Reinforcement Learning

Srinivas Venkattaramanujam

Eric Crawford

Thang Doan

Goal-conditioned policies are used in order to break down complex reinforcement learning (RL) problems by using subgoals, which can be defin… (voir plus)ed either in state space or in a latent feature space. This can increase the efficiency of learning by using a curriculum, and also enables simultaneous learning and generalization across goals. A crucial requirement of goal-conditioned policies is to be able to determine whether the goal has been achieved. Having a notion of distance to a goal is thus a crucial component of this approach. However, it is not straightforward to come up with an appropriate distance, and in some tasks, the goal space may not even be known a priori. In this work we learn a distance-to-goal estimate which is computed in terms of the number of actions that would need to be carried out in a self-supervised approach. Our method solves complex tasks without prior domain knowledge in the online setting in three different scenarios in the context of goal-conditioned policies a) the goal space is the same as the state space b) the goal space is given but an appropriate distance is unknown and c) the state space is accessible, but only a subset of the state space represents desired goals, and this subset is known a priori. We also propose a goal-generation mechanism as a secondary contribution.

2019-07-04

ArXiv (prépublication)

Singular Value Automata and Approximate Minimization

Borja Balle

Prakash Panangaden

The present paper uses spectral theory of linear operators to construct approximately minimal realizations of weighted languages. Our new co… (voir plus)ntributions are: (i) a new algorithm for the SVD decomposition of infinite Hankel matrices based on their representation in terms of weighted automata, (ii) a new canonical form for weighted automata arising from the SVD of its corresponding Hankel matrix and (iii) an algorithm to construct approximate minimizations of given weighted automata by truncating the canonical form. We give bounds on the quality of our approximation.

2019-05-26

Mathematical Structures in Computer Science (publié)

Off-Policy Deep Reinforcement Learning without Exploration

Scott Fujimoto

David Meger

Many practical applications of reinforcement learning constrain agents to learn from a fixed batch of data which has already been gathered, … (voir plus)without offering further possibility for data collection. In this paper, we demonstrate that due to errors introduced by extrapolation, standard off-policy deep reinforcement learning algorithms, such as DQN and DDPG, are incapable of learning with data uncorrelated to the distribution under the current policy, making them ineffective for this fixed batch setting. We introduce a novel class of off-policy algorithms, batch-constrained reinforcement learning, which restricts the action space in order to force the agent towards behaving close to on-policy with respect to a subset of the given data. We present the first continuous control deep reinforcement learning algorithm which can learn effectively from arbitrary, fixed batch data, and empirically demonstrate the quality of its behavior in several tasks.

2019-05-23

Proceedings of the 36th International Conference on Machine Learning (publié)

proceedings.mlr.press

Per-Decision Option Discounting

Anna Harutyunyan

Peter Vrancx

Philippe Hamel

Ann Nowé

In order to solve complex problems an agent must be able to reason over a sufficiently long horizon. Temporal abstraction, commonly modeled … (voir plus)through options, offers the ability to reason at many timescales, but the horizon length is still determined by the discount factor of the underlying Markov Decision Process. We propose a modification to the options framework that naturally scales the agent’s horizon with option length. We show that the proposed option-step discount controls a bias-variance trade-off, with larger discounts (counter-intuitively) leading to less estimation variance.

2019-05-23

Proceedings of the 36th International Conference on Machine Learning (publié)

proceedings.mlr.press

Building Knowledge for AI Agents with Reinforcement Learning

Reinforcement learning allows autonomous agents to learn how to act in a stochastic, unknown environment, with which they can interact. Deep… (voir plus) reinforcement learning, in particular, has achieved great success in well-defined application domains, such as Go or chess, in which an agent has to learn how to act and there is a clear success criterion. In this talk, I will focus on the potential role of reinforcement learning as a tool for building knowledge representations in AI agents whose goal is to perform continual learning. I will examine a key concept in reinforcement learning, the value function, and discuss its generalization to support various forms of predictive knowledge. I will also discuss the role of temporally extended actions, and their associated predictive models, in learning procedural knowledge. Finally, I will discuss the challenge of how to evaluate reinforcement learning agents whose goal is not just to control their environment, but also to build knowledge about their world.

2019-05-07

International Joint Conference on Autonomous Agents and Multiagent Systems (publié)

Connecting Weighted Automata and Recurrent Neural Networks through Spectral Learning

Guillaume Rabusseau

Tianyu Li

In this paper, we unravel a fundamental connection between weighted finite automata~(WFAs) and second-order recurrent neural networks~(2-RNN… (voir plus)s): in the case of sequences of discrete symbols, WFAs and 2-RNNs with linear activation functions are expressively equivalent. Motivated by this result, we build upon a recent extension of the spectral learning algorithm to vector-valued WFAs and propose the first provable learning algorithm for linear 2-RNNs defined over sequences of continuous input vectors. This algorithm relies on estimating low rank sub-blocks of the so-called Hankel tensor, from which the parameters of a linear 2-RNN can be provably recovered. The performances of the proposed method are assessed in a simulation study.

2019-04-10

Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics (publié)

proceedings.mlr.press

Learning Modular Safe Policies in the Bandit Setting with Application to Adaptive Clinical Trials

Hossein Aboutalebi

Tibor Schuster

The stochastic multi-armed bandit problem is a well-known model for studying the exploration-exploitation trade-off. It has significant poss… (voir plus)ible applications in adaptive clinical trials, which allow for dynamic changes in the treatment allocation probabilities of patients. However, most bandit learning algorithms are designed with the goal of minimizing the expected regret. While this approach is useful in many areas, in clinical trials, it can be sensitive to outlier data, especially when the sample size is small. In this paper, we define and study a new robustness criterion for bandit problems. Specifically, we consider optimizing a function of the distribution of returns as a regret measure. This provides practitioners more flexibility to define an appropriate regret measure. The learning algorithm we propose to solve this type of problem is a modification of the BESA algorithm [Baransi et al., 2014], which considers a more general version of regret. We present a regret bound for our approach and evaluate it empirically both on synthetic problems as well as on a dataset from the clinical trial literature. Our approach compares favorably to a suite of standard bandit algorithms.

2019-03-03

ArXiv (prépublication)

Prediction of Progression in Multiple Sclerosis Patients

Adrian Tousignant

Paul Lemaitre

Douglas Arnold

Tal Arbel

We present the first automatic end-to-end deep learning framework for the prediction of future patient disability progression (one year from… (voir plus) baseline) based on multi-modal brain Magnetic Resonance Images (MRI) of patients with Multiple Sclerosis (MS). The model uses parallel convolutional pathways, an idea introduced by the popular Inception net and is trained and tested on two large proprietary, multi-scanner, multi-center, clinical trial datasets of patients with Relapsing-Remitting Multiple Sclerosis (RRMS). Experiments on 465 patients on the placebo arms of the trials indicate that the model can accurately predict future disease progression, measured by a sustained increase in the extended disability status scale (EDSS) score over time. Using only the multi-modal MRI provided at baseline, the model achieves an AUC of 0.66 +- 0.055. However, when supplemental lesion label masks are provided as inputs as well, the AUC increases to 0.701 +- 0.027. Furthermore, we demonstrate that uncertainty estimates based on Monte Carlo dropout sample variance correlate with errors made by the model. Clinicians provided with the predictions computed by the model can therefore use the associated uncertainty estimates to assess which scans require further examination.

2019-02-27

MIDL.io/2019/Conference (poster)

openreview.net

The Termination Critic

Anna Harutyunyan

Will Dabney

Diana Borsa

Nicolas Heess

Remi Munos

In this work, we consider the problem of autonomously discovering behavioral abstractions, or options, for reinforcement learning agents. We… (voir plus) propose an algorithm that focuses on the termination function, as opposed to - as is common - the policy. The termination function is usually trained to optimize a control objective: an option ought to terminate if another has better value. We offer a different, information-theoretic perspective, and propose that terminations should focus instead on the compressibility of the option’s encoding - arguably a key reason for using abstractions.To achieve this algorithmically, we leverage the classical options framework, and learn the option transition model as a “critic” for the termination function. Using this model, we derive gradients that optimize the desired criteria. We show that the resulting options are non-trivial, intuitively meaningful, and useful for learning.

2019-02-25

ArXiv (prépublication)