AI RESEARCH
Adversarial Flow Matching for Imperceptible Attacks on End-to-End Autonomous Driving
arXiv CS.CV
•
ArXi:2605.00880v1 Announce Type: new Autonomous driving (AD) is evolving towards end-to-end (E2E) frameworks through two primary paradigms: monolithic models exemplified by Vision-Language-Action (VLA), and specialized modular architectures. Despite their divergent designs, both paradigms increasingly rely on Transformer backbones for complex reasoning, potentially causing a shared vulnerability: visually imperceptible perturbations can manipulate E2E AD models into hazardous maneuvers by targeting the Transformer module.