Scene-Adaptive VLA: Efficient Autonomous Driving via Dynamic Layer Routing

Key points

• However, the massive parameter scale of VLAs strains onboard computational resources during real-time inference.1
• Current research on efficient VLA-based driving focuses on sequential adaptation (e.g., Chain-of-Thought and token adjustment), leaving structural adaptation underexplored despite parameter redundancy across dynamic driving scenes.1
• To address this gap, we propose Scene-Adaptive VLA, an efficient framework leveraging dynamic layer routing to adjust active parameters based on real-time scenes.1
• Extensive evaluations on the Bench2Drive benchmark show that when adapted to a state-of-the-art VLA model, our approach significantly reduces inference-time computational overhead by 53.5% with negligible degradation in driving and language capabilities.1

Sources (1)

• [1]Scene-Adaptive VLA: Efficient Autonomous Driving via Dynamic Layer Routing
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 17, 09:53 AM

  Vision-Language-Action (VLA) models offer a promising paradigm for autonomous driving.

  However, the massive parameter scale of VLAs strains onboard computational resources during real-time inference.

Extractive summary: sentences quoted from the sources.