Adversarial Attacks on Multi-task Visual Perception for Autonomous Driving
This addresses safety concerns in autonomous driving by highlighting vulnerabilities in multi-task DNNs, but it is incremental as it applies known attack methods to a new multi-task context.
The paper investigates adversarial attacks on a multi-task visual perception network for autonomous driving, applying both white and black box attacks across tasks like distance estimation and object detection, and examines the effects on all tasks and a simple defense method.
Deep neural networks (DNNs) have accomplished impressive success in various applications, including autonomous driving perception tasks, in recent years. On the other hand, current deep neural networks are easily fooled by adversarial attacks. This vulnerability raises significant concerns, particularly in safety-critical applications. As a result, research into attacking and defending DNNs has gained much coverage. In this work, detailed adversarial attacks are applied on a diverse multi-task visual perception deep network across distance estimation, semantic segmentation, motion detection, and object detection. The experiments consider both white and black box attacks for targeted and un-targeted cases, while attacking a task and inspecting the effect on all the others, in addition to inspecting the effect of applying a simple defense method. We conclude this paper by comparing and discussing the experimental results, proposing insights and future work. The visualizations of the attacks are available at https://youtu.be/6AixN90budY.