适应的盲区：量化和缓解在微调驾驶模型中的遗忘
The Blind Spot of Adaptation: Quantifying and Mitigating Forgetting in Fine-tuned Driving Models
AI总结：
本文首次系统研究驾驶模型中灾难性遗忘问题，提出Drive Expert Adapter框架，通过提示空间适应提升驾驶性能并保留通用能力。
摘要：
将视觉语言模型（VLMs）整合到自动驾驶中有望解决长尾场景，但这一范式面临关键且未解决的挑战：灾难性遗忘。微调过程同时侵蚀了模型的预训练世界知识，形成自我毁灭的悖论。本文首次系统研究此现象，提出包含180万场景的大规模数据集，设计首个用于量化自动驾驶中灾难性遗忘的基准。分析表明现有方法存在显著知识退化。为此，我们提出Drive Expert Adapter（DEA）框架，通过将适应从权重空间转移到提示空间，动态根据场景特征路由推理，提升驾驶任务性能而不破坏模型基础参数。实验表明，我们的方法在驾驶任务上取得最优结果，并有效缓解灾难性遗忘，保留使VLMs成为自动驾驶系统变革力量的必要通用能力。数据和模型在FidelityDrivingBench发布。
英文摘要：
The integration of Vision-Language Models (VLMs) into autonomous driving promises to solve long-tail scenarios, but this paradigm faces the critical and unaddressed challenge of catastrophic forgetting. The very fine-tuning process used to adapt these models to driving-specific data simultaneously erodes their invaluable pre-trained world knowledge, creating a self-defeating paradox that undermines the core reason for their use. This paper provides the first systematic investigation into this phenomenon. We introduce a new large-scale dataset of 180K scenes, which enables the first-ever benchmark specifically designed to quantify catastrophic forgetting in autonomous driving. Our analysis reveals that existing methods suffer from significant knowledge degradation. To address this, we propose the Drive Expert Adapter (DEA), a novel framework that circumvents this trade-off by shifting adaptation from the weight space to the prompt space. DEA dynamically routes inference through different knowledge experts based on scene-specific cues, enhancing driving-task performance without corrupting the model's foundational parameters. Extensive experiments demonstrate that our approach not only achieves state-of-the-art results on driving tasks but also effectively mitigates catastrophic forgetting, preserving the essential generalization capabilities that make VLMs a transformative force for autonomous systems. Data and model are released at FidelityDrivingBench.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
