Doina Precup

Kushal Arora

PhD - McGill University

Principal supervisor :

Sumana Basu

PhD - McGill University

Co-supervisor :

Adriana Romero Soriano

PhD - McGill University

Élodie Côté-Gauthier Coté-Gauthier

Lynn Cherif

Master's Research - McGill University

Raymond Chua

PhD - McGill University

Co-supervisor :

PhD - McGill University

Principal supervisor :

David Meger

Jonathan Colaço Carr

Master's Research - McGill University

Principal supervisor :

Prakash Panangaden

Research Intern - McGill University

Nathan de Lara

Research Intern - McGill University

Manuel Del Verme

PhD - McGill University

Jesse Farebrother

PhD - McGill University

Principal supervisor :

Marc Gendron-Bellemare

PhD - McGill University

Co-supervisor :

David Meger

Will Hua

Master's Research - McGill University

Co-supervisor :

Guy Wolf

Howard Huang

PhD - McGill University

Haque Ishfaq

PhD - McGill University

PhD - McGill University

Mohammad Sami Nur Islam Islam

Research Intern - McGill University

Arushi Jain

PhD - McGill University

Master's Research - Université de Montréal

Principal supervisor :

PhD - McGill University

Postdoctorate - McGill University

Elaine Lau

Master's Research - McGill University

Jonathan Lebensold

PhD - McGill University

Sitao Luan

PhD - McGill University

Ray Luo

PhD - McGill University

Principal supervisor :

G McCracken

PhD - McGill University

Jaume Minano Masip

PhD - McGill University

Shahrad Mohammadzadeh

Master's Research - McGill University

Principal supervisor :

Gabriela Moisescu-Pareja

Master's Research - McGill University

Girdhar Neil Girdhar

Collaborating researcher - McGill University

Padideh Nouri

Master's Research - Université de Montréal

Charles Onu

PhD - McGill University

PhD - McGill University

Co-supervisor :

Nate Rahn

PhD - McGill University

Principal supervisor :

Marc Gendron-Bellemare

Janarthanan Rajendran

Collaborating Alumni - Université de Montréal

Principal supervisor :

Sarath Chandar

Sahand Rezaei-Shoshtari

PhD - McGill University

Co-supervisor :

PhD - McGill University

Co-supervisor :

Blake Richards

samiemandana@gmail.com

Shishir Sharma

PhD - McGill University

Nishanth Anand Vemgal

PhD - McGill University

Co-supervisor :

Master's Research - McGill University

PhD - McGill University

Priyesh Vijayan

PhD - McGill University

Co-supervisor :

Samira Ebrahimi Kahou

Research Intern - McGill University

Keyu Wang

Research Intern - McGill University

Skipper: Combining Spatial and Temporal Abstraction for Better Generalization

Steve Wen

Undergraduate - McGill University

Shuyuan Zhang

PhD - McGill University

Harry Zhao

PhD - McGill University

Co-supervisor :

Blog Posts

Generic thumbnail for Mila Blog articles.

February 22, 2024

Mingde Harry Zhao

Safa Alver

Harm van Seijen

Romain Laroche

Doina Precup

Yoshua Bengio

Read the article

Publications

Soft Condorcet Optimization for Ranking of General Agents

Marc Lanctot

Kate Larson

Michael Kaisers

Quentin Berthet

Ian Gemp

Manfred Diaz

Roberto-Rafael Maura-Rivero

Yoram Bachrach

Anna Koop

A common way to drive progress of AI models and agents is to compare their performance on standardized benchmarks. Comparing the performance… (see more) of general agents requires aggregating their individual performances across a potentially wide variety of different tasks. In this paper, we describe a novel ranking scheme inspired by social choice frameworks, called Soft Condorcet Optimization (SCO), to compute the optimal ranking of agents: the one that makes the fewest mistakes in predicting the agent comparisons in the evaluation data. This optimal ranking is the maximum likelihood estimate when evaluation data (which we view as votes) are interpreted as noisy samples from a ground truth ranking, a solution to Condorcet's original voting system criteria. SCO ratings are maximal for Condorcet winners when they exist, which we show is not necessarily true for the classical rating system Elo. We propose three optimization algorithms to compute SCO ratings and evaluate their empirical performance. When serving as an approximation to the Kemeny-Young voting method, SCO rankings are on average 0 to 0.043 away from the optimal ranking in normalized Kendall-tau distance across 865 preference profiles from the PrefLib open ranking archive. In a simulated noisy tournament setting, SCO achieves accurate approximations to the ground truth ranking and the best among several baselines when 59\% or more of the preference data is missing. Finally, SCO ranking provides the best approximation to the optimal ranking, measured on held-out test sets, in a problem containing 52,958 human players across 31,049 games of the classic seven-player game of Diplomacy.

2024-10-31

ArXiv (preprint)

Identifying and Addressing Delusions for Target-Directed Decision-Making

Harry Zhao

Mingde Zhao

Tristan Sylvain

Yoshua Bengio

We are interested in target-directed agents, which produce targets during decision-time planning, to guide their behaviors and achieve bette… (see more)r generalization during evaluation. Improper training of these agents can result in delusions: the agent may come to hold false beliefs about the targets, which cannot be properly rejected, leading to unwanted behaviors and damaging out-of-distribution generalization. We identify different types of delusions by using intuitive examples in carefully controlled environments, and investigate their causes. We demonstrate how delusions can be addressed for agents trained by hindsight relabeling, a mainstream approach in for training target-directed RL agents. We validate empirically the effectiveness of the proposed solutions in correcting delusional behaviors and improving out-of-distribution generalization.

2024-10-12

NeurIPS.cc/2024/Workshop/SafeGenAi (poster)

Mitigating Downstream Model Risks via Model Provenance

Keyu Wang

Abdullah Norozi Iranzad

Scott Schaffter

Jonathan Lebensold

Research and industry are rapidly advancing the innovation and adoption of foundation model-based systems, yet the tools for managing these … (see more)models have not kept pace. Understanding the provenance and lineage of models is critical for researchers, industry, regulators, and public trust. While model cards and system cards were designed to provide transparency, they fall short in key areas: tracing model genealogy, enabling machine readability, offering reliable centralized management systems, and fostering consistent creation incentives. This challenge mirrors issues in software supply chain security, but AI/ML remains at an earlier stage of maturity. Addressing these gaps requires industry-standard tooling that can be adopted by foundation model publishers, open-source model innovators, and major distribution platforms. We propose a machine-readable model specification format to simplify the creation of model records, thereby reducing error-prone human effort, notably when a new model inherits most of its design from a foundation model. Our solution explicitly traces relationships between upstream and downstream models, enhancing transparency and traceability across the model lifecycle. To facilitate the adoption, we introduce the unified model record (UMR) repository , a semantically versioned system that automates the publication of model records to multiple formats (PDF, HTML, LaTeX) and provides a hosted web interface (https://modelrecord.com/). This proof of concept aims to set a new standard for managing foundation models, bridging the gap between innovation and responsible model management.

2024-10-03

ArXiv (preprint)

EnzymeFlow: Generating Reaction-specific Enzyme Catalytic Pockets through Flow Matching and Co-Evolutionary Dynamics

Chenqing Hua

Yong Liu

Dinghuai Zhang

Odin Zhang

Sitao Luan

Kevin K. Yang

Guy Wolf

Shuangjia Zheng

2024-10-01

ArXiv (preprint)

Reactzyme: A Benchmark for Enzyme-Reaction Prediction

Chenqing Hua

Bozitao Zhong

Sitao Luan

Liang Hong

Guy Wolf

Shuangjia Zheng

2024-09-26

NeurIPS.cc/2024/Datasets_and_Benchmarks_Track (poster)

Adaptive Exploration for Data-Efficient General Value Function Evaluations

Arushi Jain

Josiah P. Hanna

General Value Functions (GVFs) (Sutton et al, 2011) are an established way to represent predictive knowledge in reinforcement learning. Each… (see more) GVF computes the expected return for a given policy, based on a unique pseudo-reward. Multiple GVFs can be estimated in parallel using off-policy learning from a single stream of data, often sourced from a fixed behavior policy or pre-collected dataset. This leaves an open question: how can behavior policy be chosen for data-efficient GVF learning? To address this gap, we propose GVFExplorer, which aims at learning a behavior policy that efficiently gathers data for evaluating multiple GVFs in parallel. This behavior policy selects actions in proportion to the total variance in the return across all GVFs, reducing the number of environmental interactions. To enable accurate variance estimation, we use a recently proposed temporal-difference-style variance estimator. We prove that each behavior policy update reduces the mean squared error in the summed predictions over all GVFs. We empirically demonstrate our method's performance in both tabular representations and nonlinear function approximation.

2024-09-25

NeurIPS.cc/2024/Conference (poster)

Efficient Reinforcement Learning by Discovering Neural Pathways

Samin Yeasar Arnob

Riyasat Ohib

Sergey Plis

Amy Zhang

Alessandro Sordoni

Reinforcement learning (RL) algorithms have been very successful at tackling complex control problems, such as AlphaGo or fusion control. Ho… (see more)wever, current research mainly emphasizes solution quality, often achieved by using large models trained on large amounts of data, and does not account for the financial, environmental, and societal costs associated with developing and deploying such models. Modern neural networks are often overparameterized and a significant number of parameters can be pruned without meaningful loss in performance, resulting in more efficient use of the model's capacity lottery ticket. We present a methodology for identifying sub-networks within a larger network in reinforcement learning (RL). We call such sub-networks, neural pathways. We show empirically that even very small learned sub-networks, using less than 5% of the large network's parameters, can provide very good quality solutions. We also demonstrate the training of multiple pathways within the same networks in a multitask setup, where each pathway is encouraged to tackle a separate task. We evaluate empirically our approach on several continuous control tasks, in both online and offline training

2024-09-25

NeurIPS.cc/2024/Conference (poster)

Learning Successor Features the Simple Way

Raymond Chua

Arna Ghosh

Christos Kaplanis

Blake Richards

In Deep Reinforcement Learning (RL), it is a challenge to learn representations that do not exhibit catastrophic forgetting or interference … (see more)in non-stationary environments. Successor Features (SFs) offer a potential solution to this challenge. However, canonical techniques for learning SFs from pixel-level observations often lead to representation collapse, wherein representations degenerate and fail to capture meaningful variations in the data. More recent methods for learning SFs can avoid representation collapse, but they often involve complex losses and multiple learning phases, reducing their efficiency. We introduce a novel, simple method for learning SFs directly from pixels. Our approach uses a combination of a Temporal-difference (TD) loss and a reward prediction loss, which together capture the basic mathematical definition of SFs. We show that our approach matches or outperforms existing SF learning techniques in both 2D (Minigrid), 3D (Miniworld) mazes and Mujoco, for both single and continual learning scenarios. As well, our technique is efficient, and can reach higher levels of performance in less time than other approaches. Our work provides a new, streamlined technique for learning SFs directly from pixel observations, with no pretraining required.

2024-09-25

NeurIPS.cc/2024/Conference (poster)

Offline Multitask Representation Learning for Reinforcement Learning

Haque Ishfaq

Thanh Nguyen-Tang

Songtao Feng

Raman Arora

Mengdi Wang

Ming Yin

2024-09-25

NeurIPS.cc/2024/Conference (poster)

Parseval Regularization for Continual Reinforcement Learning

Wesley Chung

Lynn Cherif

David Meger

2024-09-25

NeurIPS.cc/2024/Conference (poster)

QGFN: Controllable Greediness with Action Values

Elaine Lau

Stephen Zhewen Lu

Ling Pan

Emmanuel Bengio

Generative Flow Networks (GFlowNets; GFNs) are a family of energy-based generative methods for combinatorial objects, capable of generating … (see more)diverse and high-utility samples. However, consistently biasing GFNs towards producing high-utility samples is non-trivial. In this work, we leverage connections between GFNs and reinforcement learning (RL) and propose to combine the GFN policy with an action-value estimate,

2024-09-25

NeurIPS.cc/2024/Conference (poster)

Training Language Models to Self-Correct via Reinforcement Learning

Aviral Kumar

Vincent Zhuang

Rishabh Agarwal

Yi Su

John D Co-Reyes

Avi Singh

Kate Baumli

Shariq N Iqbal

Colton Bishop

Rebecca Roelofs

Lei M Zhang

Kay McKinney

Disha Shrivastava

Cosmin Paduraru

George Tucker

Feryal Behbahani

Aleksandra Faust

Self-correction is a highly desirable capability of large language models (LLMs), yet it has consistently been found to be largely ineffecti… (see more)ve in modern LLMs. Existing approaches for training self-correction either require multiple models or rely on a more capable model or other forms of supervision. To this end, we develop a multi-turn online reinforcement learning (RL) approach, SCoRe, that significantly improves an LLM's self-correction ability using entirely self-generated data. To build SCoRe, we first show that variants of supervised fine-tuning (SFT) on offline model-generated correction traces are insufficient for instilling self-correction behavior. In particular, we observe that training via SFT either suffers from a distribution mismatch between the training data and the model's own responses or implicitly prefers only a certain mode of correction behavior that is often not effective at test time. SCoRe addresses these challenges by training under the model's own distribution of self-generated correction traces and using appropriate regularization to steer the learning process into learning a self-correction strategy that is effective at test time as opposed to simply fitting high-reward responses for a given prompt. This regularization prescribes running a first phase of RL on a base model to generate a policy initialization that is less susceptible to collapse and then using a reward bonus to amplify self-correction during training. When applied to Gemini 1.0 Pro and 1.5 Flash models, we find that SCoRe achieves state-of-the-art self-correction performance, improving the base models' self-correction by 15.6% and 9.1% respectively on the MATH and HumanEval benchmarks.

2024-09-19

ArXiv (preprint)