LEAD: Minimizing Learner-Expert Asymmetry in End-to-End Driving

TL;DR

This paper identifies key asymmetries between expert demonstrations and student observations in simulation driving, proposing interventions that significantly improve imitation learning performance in CARLA and real-world benchmarks.

Contribution

It introduces practical methods to reduce expert-student asymmetries, leading to state-of-the-art results in simulation and real-world driving benchmarks.

Findings

Achieved 95 DS on CARLA Bench2Drive

More than doubled performance on Longest6~v2 and Town13

Gained consistent improvements on NAVSIM and Waymo benchmarks

Abstract

Simulators can generate virtually unlimited driving data, yet imitation learning policies in simulation still struggle to achieve robust closed-loop performance. Motivated by this gap, we empirically study how misalignment between privileged expert demonstrations and sensor-based student observations can limit the effectiveness of imitation learning. More precisely, experts have significantly higher visibility (e.g., ignoring occlusions) and far lower uncertainty (e.g., knowing other vehicles' actions), making them difficult to imitate reliably. Furthermore, navigational intent (i.e., the route to follow) is under-specified in student models at test time via only a single target point. We demonstrate that these asymmetries can measurably limit driving performance in CARLA and offer practical interventions to address them. After careful modifications to narrow the gaps between expert and student, our TransFuser v6 (TFv6) student policy achieves a new state of the art on all major publicly available CARLA closed-loop benchmarks, reaching 95 DS on Bench2Drive and more than doubling prior performances on Longest6~v2 and Town13. Additionally, by integrating perception supervision from our dataset into a shared sim-to-real pipeline, we show consistent gains on the NAVSIM and Waymo Vision-Based End-to-End driving benchmarks. Our code, data, and models are publicly available at https://github.com/autonomousvision/lead.