Searching Realistic-Looking Adversarial Objects For Autonomous Driving Systems
This work addresses the real-world applicability of adversarial attacks on self-driving systems, though it builds incrementally on prior research.
The paper tackles the problem of generating realistic-looking adversarial objects for autonomous driving systems by proposing a modified gradient-based texture optimization method that incorporates a 'Judge' agent to assess realism, finding that fine-tuning vision-language models or pretraining neurosymbolic systems show promise for future research.
Numerous studies on adversarial attacks targeting self-driving policies fail to incorporate realistic-looking adversarial objects, limiting real-world applicability. Building upon prior research that facilitated the transition of adversarial objects from simulations to practical applications, this paper discusses a modified gradient-based texture optimization method to discover realistic-looking adversarial objects. While retaining the core architecture and techniques of the prior research, the proposed addition involves an entity termed the 'Judge'. This agent assesses the texture of a rendered object, assigning a probability score reflecting its realism. This score is integrated into the loss function to encourage the NeRF object renderer to concurrently learn realistic and adversarial textures. The paper analyzes four strategies for developing a robust 'Judge': 1) Leveraging cutting-edge vision-language models. 2) Fine-tuning open-sourced vision-language models. 3) Pretraining neurosymbolic systems. 4) Utilizing traditional image processing techniques. Our findings indicate that strategies 1) and 4) yield less reliable outcomes, pointing towards strategies 2) or 3) as more promising directions for future research.