{"ID":2861379,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.01795","arxiv_id":"2510.01795","title":"Nav-EE: Navigation-Guided Early Exiting for Efficient Vision-Language Models in Autonomous Driving","abstract":"Vision-Language Models (VLMs) are increasingly applied in autonomous driving for unified perception and reasoning, but high inference latency hinders real-time deployment. Early-exit reduces latency by terminating inference at intermediate layers, yet its task-dependent nature limits generalization across diverse scenarios. We observe that this limitation aligns with autonomous driving: navigation systems can anticipate upcoming contexts (e.g., intersections, traffic lights), indicating which tasks will be required. We propose Nav-EE, a navigation-guided early-exit framework that precomputes task-specific exit layers offline and dynamically applies them online based on navigation priors. Experiments on CODA, Waymo, and BOSCH show that Nav-EE achieves accuracy comparable to full inference while reducing latency by up to 63.9%. Real-vehicle integration with Autoware Universe further demonstrates reduced inference latency (600ms to 300ms), supporting faster decision-making in complex scenarios. These results suggest that coupling navigation foresight with early-exit offers a viable path toward efficient deployment of large models in autonomous systems. Code and data are available at our anonymous repository: https://anonymous.4open.science/r/Nav-EE-BBC4","short_abstract":"Vision-Language Models (VLMs) are increasingly applied in autonomous driving for unified perception and reasoning, but high inference latency hinders real-time deployment. Early-exit reduces latency by terminating inference at intermediate layers, yet its task-dependent nature limits generalization across diverse scena...","url_abs":"https://arxiv.org/abs/2510.01795","url_pdf":"https://arxiv.org/pdf/2510.01795v2","authors":"[\"Haibo Hu\",\"Lianming Huang\",\"Xinyu Wang\",\"Yufei Cui\",\"Shangyu Wu\",\"Nan Guan\",\"Chun Jason Xue\"]","published":"2025-10-02T08:37:58Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"cs.AI\"]","methods":"[\"Language Model\"]","project_urls":"[\"https://anonymous.4open.science/r/Nav-EE-BBC4\"]","has_code":false}