Liam Paull

Poutine: Vision-Language-Trajectory Pre-Training and Reinforcement Learning Post-Training Enable Robust End-to-End Autonomous Driving

Luke Rowe

Rodrigue De Schaetzen

Roger Girgis

We present Poutine, a 3B-parameter vision-language model (VLM) tailored for end-to-end autonomous driving in long-tail driving scenarios. Po… (voir plus)utine is trained in two stages. To obtain strong base driving capabilities, we train Poutine-Base in a self-supervised vision-language-trajectory (VLT) next-token prediction fashion on 83 hours of CoVLA nominal driving and 11 hours of Waymo long-tail driving. Accompanying language annotations are auto-generated with a 72B-parameter VLM. Poutine is obtained by fine-tuning Poutine-Base with Group Relative Policy Optimization (GRPO) using less than 500 preference-labeled frames from the Waymo validation set. We show that both VLT pretraining and RL fine-tuning are critical to attain strong driving performance in the long-tail. Poutine-Base achieves a rater-feedback score (RFS) of 8.12 on the validation set, nearly matching Waymo's expert ground-truth RFS. The final Poutine model achieves an RFS of 7.99 on the official Waymo test set, placing 1st in the 2025 Waymo Vision-Based End-to-End Driving Challenge by a significant margin. These results highlight the promise of scalable VLT pre-training and lightweight RL fine-tuning to enable robust and generalizable autonomy.

2025-06-12

ArXiv (prépublication)

Poutine: Vision-Language-Trajectory Pre-Training and Reinforcement Learning Post-Training Enable Robust End-to-End Autonomous Driving

Luke Rowe

Rodrigue De Schaetzen

Roger Girgis

We present Poutine, a 3B-parameter vision-language model (VLM) tailored for end-to-end autonomous driving in long-tail driving scenarios. Po… (voir plus)utine is trained in two stages. To obtain strong base driving capabilities, we train Poutine-Base in a self-supervised vision-language-trajectory (VLT) next-token prediction fashion on 83 hours of CoVLA nominal driving and 11 hours of Waymo long-tail driving. Accompanying language annotations are auto-generated with a 72B-parameter VLM. Poutine is obtained by fine-tuning Poutine-Base with Group Relative Policy Optimization (GRPO) using less than 500 preference-labeled frames from the Waymo validation set. We show that both VLT pretraining and RL fine-tuning are critical to attain strong driving performance in the long-tail. Poutine-Base achieves a rater-feedback score (RFS) of 8.12 on the validation set, nearly matching Waymo's expert ground-truth RFS. The final Poutine model achieves an RFS of 7.99 on the official Waymo test set, placing 1st in the 2025 Waymo Vision-Based End-to-End Driving Challenge by a significant margin. These results highlight the promise of scalable VLT pre-training and lightweight RL fine-tuning to enable robust and generalizable autonomy.

2025-06-12

ArXiv (prépublication)

Ctrl-Crash: Controllable Diffusion for Realistic Car Crashes

Anthony Gosselin

Ge Ya Luo

Luis Lara

Florian Golemo

Derek Nowrouzezahrai

Alexia Jolicoeur-Martineau

Video diffusion techniques have advanced significantly in recent years; however, they struggle to generate realistic imagery of car crashes … (voir plus)due to the scarcity of accident events in most driving datasets. Improving traffic safety requires realistic and controllable accident simulations. To tackle the problem, we propose Ctrl-Crash, a controllable car crash video generation model that conditions on signals such as bounding boxes, crash types, and an initial image frame. Our approach enables counterfactual scenario generation where minor variations in input can lead to dramatically different crash outcomes. To support fine-grained control at inference time, we leverage classifier-free guidance with independently tunable scales for each conditioning signal. Ctrl-Crash achieves state-of-the-art performance across quantitative video quality metrics (e.g., FVD and JEDi) and qualitative measurements based on a human-evaluation of physical realism and video quality compared to prior diffusion-based methods.

2025-06-01

arXiv (publié)

Ctrl-Crash: Controllable Diffusion for Realistic Car Crashes

Anthony Gosselin

Ge Ya Luo

Luis Lara

Florian Golemo

Derek Nowrouzezahrai

Alexia Jolicoeur-Martineau

Video diffusion techniques have advanced significantly in recent years; however, they struggle to generate realistic imagery of car crashes … (voir plus)due to the scarcity of accident events in most driving datasets. Improving traffic safety requires realistic and controllable accident simulations. To tackle the problem, we propose Ctrl-Crash, a controllable car crash video generation model that conditions on signals such as bounding boxes, crash types, and an initial image frame. Our approach enables counterfactual scenario generation where minor variations in input can lead to dramatically different crash outcomes. To support fine-grained control at inference time, we leverage classifier-free guidance with independently tunable scales for each conditioning signal. Ctrl-Crash achieves state-of-the-art performance across quantitative video quality metrics (e.g., FVD and JEDi) and qualitative measurements based on a human-evaluation of physical realism and video quality compared to prior diffusion-based methods.

2025-05-30

ArXiv (prépublication)

OpenLex3D: A New Evaluation Benchmark for Open-Vocabulary 3D Scene Representations

Christina Kassab

Sacha Morin

Martin Büchner

Matias Mattamala

Kumaraditya Gupta

Abhinav Valada

Maurice Fallon

2025-05-12

IEEE.org/ICRA/2025/Workshop/Safe-VLM (spotlight)

Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments

Felix Heide

2025-03-28

ArXiv (prépublication)

The Harmonic Exponential Filter for Nonparametric Estimation on Motion Groups

Miguel Saavedra-Ruiz

Steven A. Parkison

Ria Arora

James Richard Forbes

Bayesian estimation is a vital tool in robotics as it allows systems to update the robot state belief using incomplete information from nois… (voir plus)y sensors. To render the state estimation problem tractable, many systems assume that the motion and measurement noise, as well as the state distribution, are unimodal and Gaussian. However, there are numerous scenarios and systems that do not comply with these assumptions. Existing nonparametric filters that are used to model multimodal distributions have drawbacks that limit their ability to represent a diverse set of distributions. This paper introduces a novel approach to nonparametric Bayesian filtering on motion groups, designed to handle multimodal distributions using harmonic exponential distributions. This approach leverages two key insights of harmonic exponential distributions: a) the product of two distributions can be expressed as the element-wise addition of their log-likelihood Fourier coefficients, and b) the convolution of two distributions can be efficiently computed as the tensor product of their Fourier coefficients. These observations enable the development of an efficient and asymptotically exact solution to the Bayes filter up to the band limit of a Fourier transform. We demonstrate our filter's performance compared with established nonparametric filtering methods across simulated and real-world localization tasks.

2025-02-01

IEEE Robotics and Automation Letters (publié)

Safety Representations for Safer Policy Learning

Kaustubh Mani

Vincent Mai

Charlie Gauthier

Annie S Chen

Samer B. Nashed

Reinforcement learning algorithms typically necessitate extensive exploration of the state space to find optimal policies. However, in safet… (voir plus)y-critical applications, the risks associated with such exploration can lead to catastrophic consequences. Existing safe exploration methods attempt to mitigate this by imposing constraints, which often result in overly conservative behaviours and inefficient learning. Heavy penalties for early constraint violations can trap agents in local optima, deterring exploration of risky yet high-reward regions of the state space. To address this, we introduce a method that explicitly learns state-conditioned safety representations. By augmenting the state features with these safety representations, our approach naturally encourages safer exploration without being excessively cautious, resulting in more efficient and safer policy learning in safety-critical scenarios. Empirical evaluations across diverse environments show that our method significantly improves task performance while reducing constraint violations during training, underscoring its effectiveness in balancing exploration with safety.

2025-01-22

ICLR.cc/2025/Conference (poster)

Safety Representations for Safer Policy Learning

Kaustubh Mani

Vincent Mai

Charlie Gauthier

Annie S Chen

Samer B. Nashed

Reinforcement learning algorithms typically necessitate extensive exploration of the state space to find optimal policies. However, in safet… (voir plus)y-critical applications, the risks associated with such exploration can lead to catastrophic consequences. Existing safe exploration methods attempt to mitigate this by imposing constraints, which often result in overly conservative behaviours and inefficient learning. Heavy penalties for early constraint violations can trap agents in local optima, deterring exploration of risky yet high-reward regions of the state space. To address this, we introduce a method that explicitly learns state-conditioned safety representations. By augmenting the state features with these safety representations, our approach naturally encourages safer exploration without being excessively cautious, resulting in more efficient and safer policy learning in safety-critical scenarios. Empirical evaluations across diverse environments show that our method significantly improves task performance while reducing constraint violations during training, underscoring its effectiveness in balancing exploration with safety.

2025-01-22

ICLR.cc/2025/Conference (poster)

Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments

Felix Heide

2025-01-01

CVPR (publié)

Rethinking Teacher-Student Curriculum Learning through the Cooperative Mechanics of Experience

Manfred Diaz

Andrea Tacchetti

Teacher-Student Curriculum Learning (TSCL) is a curriculum learning framework that draws inspiration from human cultural transmission and le… (voir plus)arning. It involves a teacher algorithm shaping the learning process of a learner algorithm by exposing it to controlled experiences. Despite its success, understanding the conditions under which TSCL is effective remains challenging. In this paper, we propose a data-centric perspective to analyze the underlying mechanics of the teacher-student interactions in TSCL. We leverage cooperative game theory to describe how the composition of the set of experiences presented by the teacher to the learner, as well as their order, influences the performance of the curriculum that is found by TSCL approaches. To do so, we demonstrate that for every TSCL problem, there exists an equivalent cooperative game, and several key components of the TSCL framework can be reinterpreted using game-theoretic principles. Through experiments covering supervised learning, reinforcement learning, and classical games, we estimate the cooperative values of experiences and use value-proportional curriculum mechanisms to construct curricula, even in cases where TSCL struggles. The framework and experimental setup we present in this work represent a novel foundation for a deeper exploration of TSCL, shedding light on its underlying mechanisms and providing insights into its broader applicability in machine learning.

2024-09-17

TMLR (accepté)