Overview
| Date | June 19, 2023 |
| Location | West 301 |
| Virtual Site | https://cvpr2023.thecvf.com/virtual/2023/workshop/18448 |
Despite the great success achieved by machine learning recently, extensive studies have shown that machine learning algorithms are vulnerable to adversarial attacks or natural distribution shifts, which has raised great concerns when deploying machine learning algorithms for real-world applications, especially in safety-critical domains such as autonomous driving (AD). While there have been significant advances in AD (e.g., perception, planning and control, etc.), the security and safety of these algorithms are often challenged by various realistic safety-critical scenarios.
In this workshop, we aim to explore and discuss recent research and summarize potential future directions for secure and safe AD algorithms. In particular, we will host different invited talks, paper submissions, panel discussions, and a safe AD competition based on our unified platform SafeBench, which is developed to integrate different types of safety-critical testing scenarios, scenario generation algorithms, and other variations such as driving routes and environments, to provide comprehensive learning and testing environment for AD algorithms.
We will bring together experts from computer vision, reinforcement learning, security, and trustworthy machine learning communities, in an attempt to highlight recent work in this area as well as to clarify the foundations of secure autonomous driving. We hope this workshop will help to chart out important directions for future work and cross-community collaborations.
We invite submissions on secure and safe autonomous driving algorithms, including (but not limited to):
- Robust perception algorithms against adversarial attack
- Physical attack and defense to autonomous driving systems
- Data poisoning and defense to autonomous driving systems
- Robust decision-making algorithms, e.g., Robust RL
- Safe decision-making and control algorithms, e.g., Safe RL
- Trajectory and behavior prediction with uncertainty
- Behavior modeling of pedestrian and vehicle
- Causal discovery and counterfactual analysis of driving behavior
- Safety verification and certification
- Neuro-symbolic approaches in autonomous driving
- Knowledge representation and reasoning in autonomous driving
- Autonomous driving datasets, simulation, evaluations, and metrics
Organizers
 |
 |
 |
 |
 |
| Chejian Xu Ph.D. Student, UIUC |
Wenhao Ding Ph.D. Student, CMU |
Haohong Lin Ph.D. Student, CMU |
Mansur Arief Ph.D. Student, CMU |
Jiawei Zhang Master Student, UIUC |
 |
 |
 |
 |
 |
| Hazem Torfah Postdoc, UCB |
Alberto Sangiovanni- Vincentelli Professor, UCB |
Sanjit A. Seshia Professor, UCB |
Ding Zhao Assistant Professor, CMU |
Bo Li Assistant Professor, UIUC |