Rowan McAllister

Staff Research Scientist

Waymo

Biography

I work on machine learning methods for autonomous vehicle planning at Waymo in California. I was fortunate to learn from Carl Rasmussen as a PhD student, and from Sergey Levine as a postdoc, focusing on probabilistic modeling for reinforcement learning. I have also enjoyed learning from Irena Koprinska, Robert Fitch, Thierry Peynot, Jeff Schneider, and Adrien Gaidon. I am passionate about autonomous vehicles, previously working at the Australian Centre for Field Robotics and UberATG (bought by Aurora), co-founded Light Blue Labs (now Wayve), managed a research team at Toyota Research Institute on prediction and planning, and organize workshops on autonomous driving.

Interests

Autonomous Vehicles
Reinforcement Learning
Bayesian Statistics
Cycling

Education

PhD in Machine Learning, 2017

University of Cambridge
MPhil in Robotics, 2012

Australian Centre for Field Robotics
BEng in Mechatronics-Space, 2009

University of Sydney
BSc in Math, CS, Physics, 2009

University of Sydney

Posts

Academia to industry

Transitioning from academic research to industry, in the US tech sector

Rowan McAllister

Aug 30, 2022 49 min read

Workshop organization guide

Advice on workshop organization at academic conferences

Rowan McAllister

May 10, 2022 46 min read

How to do research

Choosing the right research problem + how to succeed

Rowan McAllister

Oct 26, 2021 4 min read

Publications

In-Distribution Barrier Functions: Self-Supervised Policy Filters that Avoid Out-of-Distribution States

Learning-based control approaches have shown great promise in performing complex tasks directly from high-dimensional perception data …

Fernando Castaneda, Haruki Nishimura, Rowan McAllister, Koushil Sreenath, Adrien Gaidon

PDF

Best Paper Award, Learning for Dynamics and Control (LD4C) Conference 2023

RAP: Risk-Aware Prediction for Robust Planning

Robust planning in interactive scenarios requires predicting the uncertain future to make risk-aware decisions. Unfortunately, due to …

Haruki Nishimura, Jean Mercat, Blake Wulfe, Rowan McAllister, Adrien Gaidon

PDF Code Website

Oral presentation, CoRL 2022 (top 7% of submitted papers)

Is Anyone There? Learning a Planner Contingent on Perceptual Uncertainty

Robots in complex multi-agent environments should reason about the intentions of observed and currently unobserved agents. In this …

Charles Packer, Nicholas Rhinehart, Rowan McAllister, Matthew Wright, Xin Wang, Jeff He, Sergey Levine, Joseph Gonzalez

PDF Website

Control-Aware Prediction Objectives for Autonomous Driving

Autonomous vehicle software is typically structured as a modular pipeline of individual components (e.g., perception, prediction, and …

Rowan McAllister, Blake Wulfe, Jean Mercat, Logan Ellis, Sergey Levine, Adrien Gaidon

PDF Poster Video Website

Dynamics-Aware Comparison of Learned Reward Functions

The ability to learn reward functions plays an important role in enabling the deployment of intelligent agents in the real world. …

Blake Wulfe, Ashwin Balakrishna, Logan Ellis, Jean Mercat, Rowan McAllister, Adrien Gaidon

PDF Website

Spotlight presentation, ICLR 2022 (top 7% of submitted papers)

Heterogeneous-Agent Trajectory Forecasting Incorporating Class Uncertainty

Reasoning about the future behavior of other agents is critical to safe robot navigation. The multiplicity of plausible futures is …

Boris Ivanovic, Kuan-Hui Lee, Pavel Tokmakov, Blake Wulfe, Rowan McAllister, Adrien Gaidon, Marco Pavone

PDF

Contingencies from Observations: Tractable Contingency Planning with Learned Behavior Models

Humans have a remarkable ability to make decisions by accurately reasoning about future events, including the future behaviors and …

Nicholas Rhinehart, Jeff He, Charles Packer, Matthew Wright, Rowan McAllister, Joseph Gonzalez, Sergey Levine

PDF Code Video Website

Outcome-Driven Reinforcement Learning via Variational Inference

While reinforcement learning algorithms provide automated acquisition of optimal policies, practical application of such methods …

Tim Rudner, Vitchyr Pong, Rowan McAllister, Yarin Gal, Sergey Levine

PDF

Can Autonomous Vehicles Identify, Recover From, and Adapt to Distribution Shifts?

Out-of-training-distribution (OOD) scenarios are a common challenge of learning agents at deployment, typically leading to arbitrary …

Angelos Filos, Panagiotis Tigas, Rowan McAllister, Nicholas Rhinehart, Sergey Levine, Yarin Gal

PDF Code Slides Video Website

Learning Invariant Representations for Reinforcement Learning without Reconstruction

We study how representation learning can accelerate reinforcement learning from rich observations, such as images, without relying …

Amy Zhang, Rowan McAllister, Roberto Calandra, Yarin Gal, Sergey Levine

PDF Code Video Website

Oral presentation, ICLR 2021 (top 1.8% of submitted papers)

Model-Based Meta-Reinforcement Learning forFlight with Suspended Payloads

Transporting suspended payloads is challenging for autonomous aerial vehicles because the payload can cause significant and …

Suneel Belkhale, Rachel Li, Gregory Kahn, Rowan McAllister, Roberto Calandra, Sergey Levine

PDF Code Video DOI Website

Safety Augmented Value Estimation from Demonstrations (SAVED): Safe Deep Model-Based RL for Sparse Cost Robotic Tasks

Reinforcement learning (RL) for robotics is challenging due to the difficulty in hand-engineering a dense cost function, which can lead …

Brijen Thananjeyan, Ashwin Balakrishna, Ugo Rosolia, Felix Li, Rowan McAllister, Joseph Gonzalez, Sergey Levine, Francesco Borrelli, Ken Goldberg

PDF Code Slides Video Website

PRECOG: PREdiction Conditioned On Goals in Visual Multi-Agent Settings

Forecasting the motion of multiple interacting vehicles. When one is autonmous, conditioning on its goals helps better-predict the …

Nicholas Rhinehart, Rowan McAllister, Kris Kitani, Sergey Levine

PDF Code Dataset Poster Slides Video Website

Oral presentation, BayLearn 2019 (top 6 of 72 papers)

Invited talk, ICCV 2019 workshop on Autonomous Driving

Best Paper, ICML 2019 workshop on AI for Autonomous Driving

Robustness to Out-of-Distribution Inputs via Task-Aware Generative Uncertainty

Deep learning provides a powerful tool for machine perception when the observations resemble the training data. However, real-world …

Rowan McAllister, Gregory Kahn, Jeff Clune, Sergey Levine

PDF Dataset Website

Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models

Model-based reinforcement learning (RL) algorithms can attain excellent sample efficiency, but often lag behind the best model-free …

Kurtland Chua, Roberto Calandra, Rowan McAllister, Sergey Levine

PDF Code Video Website

Oral presentation, BayLearn 2018

Spotlight presentation, NeurIPS 2018 (top 3.5% of submitted papers)

NVIDIA Pioneer Award, NeurIPS 2018

Deep Imitative Models for Flexible Inference, Planning, and Control

Imitation Learning (IL) is an appealing approach to learn desirable autonomous behavior. However, directing IL to achieve arbitrary …

Nicholas Rhinehart, Rowan McAllister, Sergey Levine

PDF Code Video Website

Contributed talk at NeurIPS 2018 workshop on ML for Intelligent Transportation Systems (top 3 of 25 papers)

Contributed talk at NeurIPS 2018 workshop on Infer2Control (top 6 of 27 papers)

See all publications

See all workshops

Rowan McAllister

Staff Research Scientist

Waymo

Biography

Interests

Education

Posts

Academia to industry

Workshop organization guide

How to do research

Publications

In-Distribution Barrier Functions: Self-Supervised Policy Filters that Avoid Out-of-Distribution States

RAP: Risk-Aware Prediction for Robust Planning

Is Anyone There? Learning a Planner Contingent on Perceptual Uncertainty

Control-Aware Prediction Objectives for Autonomous Driving

Dynamics-Aware Comparison of Learned Reward Functions

Heterogeneous-Agent Trajectory Forecasting Incorporating Class Uncertainty

Contingencies from Observations: Tractable Contingency Planning with Learned Behavior Models

Outcome-Driven Reinforcement Learning via Variational Inference

Can Autonomous Vehicles Identify, Recover From, and Adapt to Distribution Shifts?

Learning Invariant Representations for Reinforcement Learning without Reconstruction

Model-Based Meta-Reinforcement Learning forFlight with Suspended Payloads

Safety Augmented Value Estimation from Demonstrations (SAVED): Safe Deep Model-Based RL for Sparse Cost Robotic Tasks

PRECOG: PREdiction Conditioned On Goals in Visual Multi-Agent Settings

Robustness to Out-of-Distribution Inputs via Task-Aware Generative Uncertainty

Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models

Deep Imitative Models for Flexible Inference, Planning, and Control

Workshops

Machine Learning for Autonomous Driving

Fresh Perspectives on the Future of Autonomous Driving

Machine Learning for Autonomous Driving

Autonomous Vehicle Vision

Multi-Agent Interaction and Relational Reasoning

Machine Learning for Autonomous Driving

Perception for Autonomous Driving

AI for Autonomous Driving

Interaction and Decision-Making in Autonomous-Driving