Portrait of Liam Paull

Liam Paull

Core Academic Member

paulll@mila.quebec

Canada CIFAR AI Chair

Assistant Professor, Université de Montréal, Department of Computer Science and Operations Research

Research Topics

Computer Vision

Deep Learning

Robotics

Biography

Liam Paull is an associate professor at Université de Montréal and co-leads the Montréal Robotics and Embodied AI Lab (REAL). His lab focuses on a variety of robotics problems, including building representations of the world for such applications as simultaneous localization and mapping, modelling uncertainty, and building better workflows to teach robotic agents new tasks through, for example, simulation or demonstration.

Previously, Paull was a research scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT), where he led the autonomous car project funded by the Toyota Research Institute (TRI). He completed a postdoc with the Marine Robotics Group at MIT, where he worked on Simultaneous Localization and Mapping (SLAM) for underwater robots.

His PhD from the University of New Brunswick in 2013 focused on robust and adaptive planning for underwater vehicles. He is also the co-founder and director of the Duckietown Foundation, which is dedicated to making engaging robotics learning experiences accessible to everyone.

Current Students

Master's Research - Université de Montréal

Rodrigue De Schaetzen

PhD - Université de Montréal

Moustafa Elarabi

PhD - Université de Montréal

Guillaume Gagné-Labelle

Collaborating researcher - Université de Montréal

Charlie Gauthier

PhD - Université de Montréal

Co-supervisor :

Kumaraditya Gupta

PhD

Co-supervisor :

PhD - Université de Montréal

PhD - Université de Montréal

Principal supervisor :

Collaborating researcher - Université Laval

Azalee Robitaille

Master's Research - Université de Montréal

PhD - Université de Montréal

Co-supervisor :

Miguel Angel Saavedra Ruiz

PhD - Université de Montréal

Collaborating researcher - Université de Montréal

Blog Posts

Visuel de l'Article sur la représentation du maillage non étanche de t-shirts

May 15, 2024

How to Effectively and Efficiently Represent Non-Watertight Meshes for Your T-Shirts

by

Zhen Liu

Yao Feng

Yuliang Xiu

Weiyang Liu

Liam Paull

Michael J. Black

Bernhard Scholkopf

Read the article

Sample Efficient Deep Reinforcement Learning Via Uncertainty Estimation

May 9, 2022

Sample Efficient Deep Reinforcement Learning Via Uncertainty Estimation

by

Vincent Mai

Liam Paull

Read the article

La-MAML: Look-ahead Meta-Learning for Continual Learning

November 19, 2020

La-MAML: Look-ahead Meta-Learning for Continual Learning

by

Gunshi Gupta

Liam Paull

Read the article

Publications

MeshDiffusion: Score-Based Generative 3D Mesh Modeling

Yao Feng

Michael J. Black

Derek Nowrouzezahrai

Weiyang Liu

We consider the task of generating realistic 3D shapes, which is useful for a variety of applications such as automatic scene generation and… (see more) physical simulation. Compared to other 3D representations like voxels and point clouds, meshes are more desirable in practice, because (1) they enable easy and arbitrary manipulation of shapes for relighting and simulation, and (2) they can fully leverage the power of modern graphics pipelines which are mostly optimized for meshes. Previous scalable methods for generating meshes typically rely on sub-optimal post-processing, and they tend to produce overly-smooth or noisy surfaces without fine-grained geometric details. To overcome these shortcomings, we take advantage of the graph structure of meshes and use a simple yet very effective generative modeling method to generate 3D meshes. Specifically, we represent meshes with deformable tetrahedral grids, and then train a diffusion model on this direct parametrization. We demonstrate the effectiveness of our model on multiple generative tasks.

2023-01-31

ICLR.cc/2023/Conference (notable)

Robust and Controllable Object-Centric Learning through Energy-based Models

Tong Che

Boris Ivanovic

Marco Pavone

Humans are remarkably good at understanding and reasoning about complex visual scenes. The capability to decompose low-level observations in… (see more)to discrete objects allows us to build a grounded abstract representation and identify the compositional structure of the world. Accordingly, it is a crucial step for machine learning models to be capable of inferring objects and their properties from visual scenes without explicit supervision. However, existing works on object-centric representation learning either rely on tailor-made neural network modules or strong probabilistic assumptions in the underlying generative and inference processes. In this work, we present \ours, a conceptually simple and general approach to learning object-centric representations through an energy-based model. By forming a permutation-invariant energy function using vanilla attention blocks readily available in Transformers, we can infer object-centric latent variables via gradient-based MCMC methods where permutation equivariance is automatically guaranteed. We show that \ours can be easily integrated into existing architectures and can effectively extract high-quality object-centric representations, leading to better segmentation accuracy and competitive downstream task performance. Further, empirical evaluations show that \ours's learned representations are robust against distribution shift. Finally, we demonstrate the effectiveness of \ours in systematic compositional generalization, by re-composing learned energy functions for novel scene generation and manipulation.

2023-01-31

ICLR.cc/2023/Conference (poster)

GROOD: Gradient-Aware Out-of-Distribution Detection

Mostafa Elaraby

Sabyasachi Sahoo

Yann Batiste Pequignot

Paul Novello

2022-12-31

arXiv.org (preprint)

Multi-Agent Reinforcement Learning for Fast-Timescale Demand Response of Residential Loads

Vincent Mai

Philippe Maisonneuve

Antoine Lesage-Landry

To integrate high amounts of renewable energy resources, electrical power grids must be able to cope with high amplitude, fast timescale var… (see more)iations in power generation. Frequency regulation through demand response has the potential to coordinate temporally flexible loads, such as air conditioners, to counteract these variations. Existing approaches for discrete control with dynamic constraints struggle to provide satisfactory performance for fast timescale action selection with hundreds of agents. We propose a decentralized agent trained with multi-agent proximal policy optimization with localized communication. We explore two communication frameworks: hand-engineered, or learned through targeted multi-agent communication. The resulting policies perform well and robustly for frequency regulation, and scale seamlessly to arbitrary numbers of houses for constant processing times.

2022-12-31

AAMAS (published)

NeurIPS 2022 Competition: Driving SMARTS

Amir Hossein Rasouli

R. Goebel

Matthew E. Taylor

Iuliia Kotseruba

Soheil Alizadeh

Tianpei Yang

Montgomery Alban

Florian Shkurti

Yuzheng Zhuang

Adam Ścibior

Kasra Rezaee

Animesh Garg

Jun Luo

Weinan Zhang

Xinyu Wang

Xiangshan Chen

2022-11-13

ArXiv (preprint)

Lifelong Topological Visual Navigation

Rey Reza Wiyatno

Anqi Xu

Commonly, learning-based topological navigation approaches produce a local policy while preserving some loose connectivity of the space thro… (see more)ugh a topological map. Nevertheless, spurious or missing edges in the topological graph often lead to navigation failure. In this work, we propose a sampling-based graph building method, which results in sparser graphs yet with higher navigation performance compared to baseline methods. We also propose graph maintenance strategies that eliminate spurious edges and expand the graph as needed, which improves lifelong navigation performance. Unlike controllers that learn from fixed training environments, we show that our model can be fine-tuned using only a small number of collected trajectory images from a real-world environment where the agent is deployed. We demonstrate successful navigation after fine-tuning on real-world environments, and notably show significant navigation improvements over time by applying our lifelong graph maintenance strategies.

2022-09-30

IEEE Robotics and Automation Letters (published)

Monocular Robot Navigation with Self-Supervised Pretrained Vision Transformers

Miguel Saavedra-Ruiz

In this work, we consider the problem of learning a perception model for monocular robot navigation using few annotated images. Using a Visi… (see more)on Transformer (ViT) pretrained with a label-free self-supervised method, we successfully train a coarse image segmentation model for the Duckietown environment using 70 training images. Our model performs coarse image segmentation at the 8x8 patch level, and the inference resolution can be adjusted to balance prediction granularity and real-time perception constraints. We study how best to adapt a ViT to our task and environment, and find that some lightweight architectures can yield good single-image segmentation at a usable frame rate, even on CPU. The resulting perception model is used as the backbone for a simple yet robust visual servoing agent, which we deploy on a differential drive mobile robot to perform two tasks: lane following and obstacle avoidance.

2022-06-01

2022 19th Conference on Robots and Vision (CRV) (published)

f-Cal: Aleatoric uncertainty quantification for robot perception via calibrated neural regression

Krishna Murthy

Hanju Lee

While modern deep neural networks are performant perception modules, performance (accuracy) alone is insufficient, particularly for safety-c… (see more)ritical robotic applications such as self-driving vehicles. Robot autonomy stacks also require these otherwise blackbox models to produce reliable and calibrated measures of confidence on their predictions. Existing approaches estimate uncertainty from these neural network perception stacks by modifying network architectures, inference procedure, or loss functions. However, in general, these methods lack calibration, meaning that the predictive uncertainties do not faithfully represent the true underlying uncertainties (process noise). Our key insight is that calibration is only achieved by imposing constraints across multiple examples, such as those in a mini-batch; as opposed to existing approaches which only impose constraints per-sample, often leading to overconfident (thus miscalibrated) uncertainty estimates. By enforcing the distribution of outputs of a neural network to resemble a target distribution by minimizing an

2021-12-31

International Conference on Robotics and Automation (published)

Perceptual Generative Autoencoders

Zijun Zhang

Zongpeng Li

Modern generative models are usually designed to match target distributions directly in the data space, where the intrinsic dimension of dat… (see more)a can be much lower than the ambient dimension. We argue that this discrepancy may contribute to the difficulties in training generative models. We therefore propose to map both the generated and target distributions to a latent space using the encoder of a standard autoencoder, and train the generator (or decoder) to match the target distribution in the latent space. Specifically, we enforce the consistency in both the data space and the latent space with theoretically justified data and latent reconstruction losses. The resulting generative model, which we call a perceptual generative autoencoder (PGA), is then trained with a maximum likelihood or variational autoencoder (VAE) objective. With maximum likelihood, PGAs generalize the idea of reversible generative models to unrestricted neural network architectures and arbitrary number of latent dimensions. When combined with VAEs, PGAs substantially improve over the baseline VAEs in terms of sample quality. Compared to other autoencoder-based generative models using simple priors, PGAs achieve state-of-the-art FID scores on CIFAR-10 and CelebA.

2020-11-20

Proceedings of the 37th International Conference on Machine Learning (published)

proceedings.mlr.press

Active Domain Randomization

Bhairav Mehta

Manfred Diaz

Christopher Pal

Domain randomization is a popular technique for improving domain transfer, often used in a zero-shot setting when the target domain is unkno… (see more)wn or cannot easily be used for training. In this work, we empirically examine the effects of domain randomization on agent generalization. Our experiments show that domain randomization may lead to suboptimal, high-variance policies, which we attribute to the uniform sampling of environment parameters. We propose Active Domain Randomization, a novel algorithm that learns a parameter sampling strategy. Our method looks for the most informative environment variations within the given randomization ranges by leveraging the discrepancies of policy rollouts in randomized and reference environment instances. We find that training more frequently on these instances leads to better overall agent generalization. In addition, when domain randomization and policy transfer fail, Active Domain Randomization offers more insight into the deficiencies of both the chosen parameter ranges and the learned policy, allowing for more focused debugging. Our experiments across various physics-based simulated and a real-robot task show that this enhancement leads to more robust, consistent policies.

2020-05-11

Proceedings of the Conference on Robot Learning (published)

proceedings.mlr.press

proceedings.mlr.press

Curriculum in Gradient-Based Meta-Reinforcement Learning

Bhairav Mehta

Sharath Chandra Raparthy

Christopher Pal

Gradient-based meta-learners such as Model-Agnostic Meta-Learning (MAML) have shown strong few-shot performance in supervised and reinforcem… (see more)ent learning settings. However, specifically in the case of meta-reinforcement learning (meta-RL), we can show that gradient-based meta-learners are sensitive to task distributions. With the wrong curriculum, agents suffer the effects of meta-overfitting, shallow adaptation, and adaptation instability. In this work, we begin by highlighting intriguing failure cases of gradient-based meta-RL and show that task distributions can wildly affect algorithmic outputs, stability, and performance. To address this problem, we leverage insights from recent literature on domain randomization and propose meta Active Domain Randomization (meta-ADR), which learns a curriculum of tasks for gradient-based meta-RL in a similar as ADR does for sim2real transfer. We show that this approach induces more stable policies on a variety of simulated locomotion and navigation tasks. We assess in- and out-of-distribution generalization and find that the learned task distributions, even in an unstructured task space, greatly improve the adaptation performance of MAML. Finally, we motivate the need for better benchmarking in meta-RL that prioritizes \textit{generalization} over single-task adaption performance.

2020-02-18

ArXiv (preprint)

Your GAN is Secretly an Energy-based Model and You Should Use Discriminator Driven Latent Sampling

Tong Che

Jascha Sohl-Dickstein

Hugo Larochelle

Yuan Cao

We show that the sum of the implicit generator log-density …

2019-12-31

Advances in Neural Information Processing Systems 33 (NeurIPS 2020) (published)