Scalability in Autonomous Driving

Introduction

While advances in machine learning have enabled cars to drive themselves, rolling out a self-driving service to larger areas and at scale might require even more capable and robust systems that generalize to previously unseen environments and driving conditions. To this end, we host a workshop to specifically address the issues ahead in scalability of autonomous driving technology. In particular, we will discuss questions such as how to adapt a self-driving system to diverse geographies, weather and road conditions, and to different driving norms around the world.

This workshop will explore, among others, the following topics:

  • Generalization in detection / tracking / scene understanding
  • Handling of long-tail events
  • Self-supervised / semi-supervised learning
  • Domain adaptation / transfer learning
  • Efficient data labeling
  • Multi-task learning

Invited speakers

Tentatively confirmed:

  • Raquel Urtasun, Uber ATG & University of Toronto
  • Paul Newman, Oxbotica & University of Oxford
  • Andrej Karpathy, Tesla
  • Alex Kendall, Wayve
  • Kris Kitani, CMU

Competitions

TBD.

Launch: 2020/03

Submission deadline: 2020/05

Organizers

  • Yuning Chai, Waymo
  • Henrik Kretzschmar, Waymo
  • Yin Zhou, Waymo
  • Pei Sun, Waymo
  • Lukas Neumann, University of Oxford
  • Andrea Vedaldi, University of Oxford & Facebook
  • Andreas Geiger, University of Tübingen & MPI
  • Dragomir Anguelov, Waymo