通过表征锚定与语言-动作对齐实现可泛化的VLA微调

摘要

通过行为克隆（BC）在机器人演示数据上微调预训练视觉-语言模型（VLM），已成为视觉-语言-动作（VLA）策略的标准范式。然而，BC微调会逐步覆盖支持视觉与语义泛化的预训练表征。联合训练网络图像-文本数据这一常见补救措施无法阻止该现象；它仅在独立观测上施加语言损失与动作损失，导致VLA存在语言-动作错位，而标准操作基准测试无法暴露这一问题。我们提出Anchor-Align方法，在BC基础上增加两个目标：视觉-语言锚定（Vision-Language Anchoring）从冻结的VLM副本中蒸馏逐层表征以防止表征漂移，而语言-动作对齐（Language-Action Alignment）将每个动作目标转化为离散的运动方向标签，并在同一机器人观测上联合训练语言预测与动作预测。在物理xArm7机器人上，针对两种广泛使用的VLA架构，Anchor-Align分别将真实机器人成功率从28%提升至54%，以及从37%提升至60%。在仿真规模化实验中，我们在LIBERO-PRO、LIBERO-Plus和CALVIN上分别展示了在分布外扰动、感知鲁棒性和长程控制方面的持续改进，表明保留预训练表征与高效动作学习并非根本对立。项目页面：anchoralignvla.github.io

English

Finetuning a pretrained vision-language model (VLM) on robot demonstrations via behavior cloning (BC) has become the standard recipe for vision-language-action (VLA) policies. However, BC finetuning progressively overwrites the pretrained representations that support visual and semantic generalization. Co-training on web image-text data, a common remedy, does not prevent this; it applies language and action losses to separate observations, leaving VLAs with language-action misalignment that standard manipulation benchmarks do not expose. We propose Anchor-Align, which augments BC with two objectives: Vision-Language Anchoring distills layer-wise representations from a frozen VLM copy to prevent this drift, while Language-Action Alignment converts each action target into a discrete motion-direction label and jointly trains language and action prediction on the same robot observation. On a physical xArm7 robot, across two widely used VLA architectures, Anchor-Align improves real-robot success on both (28% to 54% and 37% to 60%). At scale in simulation, we demonstrate consistent improvements on OOD perturbations, perceptual robustness, and long-horizon control across LIBERO-PRO, LIBERO-Plus, and CALVIN, respectively, suggesting that preserving pretrained representations and effective action learning are not fundamentally at odds. Project page: anchoralignvla.github.io