CVPR 2021 Workshop
Autonomous Driving: Perception, Prediction and Planning
Autonomous driving research can improve society by reducing road accidents, giving independence to those unable to drive, and inspiring younger generations towards computer vision with tangible examples of the technology clearly visible on local streets in many cities around the world. The goal of this workshop is to draw attention to the problems connecting the whole perception, prediction and planning stack.
We hope to encourage mutually informed progress on all three components, including the fundamental challenges of autonomous driving such as few-shot generalization, domain adaptation, self-supervised or semi-supervised learning, and planning friendly perception models. Furthermore, the workshop intends to foster interdisciplinary communication of researchers working on autonomous driving from both academia and industry. Finally, we intend to kick off a planning competition powered by a real-world self-driving dataset, which, to the best of our knowledge, has never been done in autonomous driving before.
What’s around the car?
What will likely happen next?
What should an AV do?
June 19th, Eastern Time (ET)
10:00 am Introduction
10:10 am Peter Ondruska ML planning & Simulation
10:30 am Dorsa Sadigh When our Human Modeling Assumptions Fail
10:50 am Kilian Q. Weinberger Persistence priors aid 3D detection with past traversals
11:10 am Raquel Urtasun Learning to Drive in Highly Dynamic Environments
11:30 am Q&A
11:45 am Poster session
12:30 pm Chen Wu Scaling the Perception and Autonomous Driving Systems via Efficient Data,Learning, and Simulation Techniques
12:50 pm Anca Dragan Flexible and safe intent-driven predictions
01:10 pm Charles R. Qi Offboard Perception for Autonomous Driving
01:30 pm Arun Venkatraman The Value Proposition of Forecasting for motion planning
01:50 pm Q&A
02:05 pm Contributed talks
02:45 pm Poster session
03:30 pm Shenlong Wang Towards Realistic and Scalable Simulation for Autonomous Driving
03:50 pm Sergey Levine Data-Driven Control: Reinforcement Learning without Trial and Error
04:10 pm Sanja Fidler TBD
04:30 pm Q&A
04:40 pm Panel discussion
05:30 pm Poster session
Call for papers
We welcome authors to submit their papers in two different formats: full-paper version (4-8 pages) and short-abstract version (2-pages). The full-paper should describe the work that has not been published or accepted recently. The short-abstract highlights the significant work that has been published or accepted recently. Please use the CVPR 2021 paper template and follow the CVPR submission guideline. Accepted papers will be posted on the website but there will not be archival proceedings.
The submission needs to be submitted to CMT system: https://cmt3.research.microsoft.com/ADP3CVPR2021.
2D/3D object detection for autonomous driving
Single/Multiple objects tracking
Motion prediction and planning for autonomous driving
Data-driven driving simulation
Reinforcement Learning and Imitation Learning for AV
Domain adaptation for autonomous driving
Sensor fusion for autonomous driving
Monocular/stereo depth estimation
Submission Open: April 1, 2021
April 30, May 18, 2021, 11:59 PM Pacific Time
May 15, May 28, 2021
Camera Ready Deadline: June 10, 2021