DySL-VLA：通过动态-静态层跳过实现高效的视觉-语言-动作模型推理以用于机器人操作
DySL-VLA: Efficient Vision-Language-Action Model Inference via Dynamic-Static Layer-Skipping for Robot Manipulation
AI总结：
本文提出DySL-VLA框架，通过动态跳过视觉-语言-动作层来降低计算成本，提升机器人操作任务的实时性能，实验表明在Calvin数据集上成功长度提升2.1%，参数减少85.7倍，速度提升3.75倍。
摘要：
视觉-语言-动作（VLA）模型在机器人任务如 manipulation 中通过融合语言模型的推理与视觉模型的3D理解展现出显著的成功。然而，其高计算成本仍然是限制实时应用的主要障碍。我们观察到任务中的动作具有不同的重要性级别：关键步骤需要高精度，而不太重要的步骤可以容忍更多的变化。基于这一见解，我们提出了DySL-VLA，一种新的框架，通过根据每个动作的重要性动态跳过VLA层来解决计算成本问题。DySL-VLA将它的层分为两种类型：信息层，始终执行；以及增量层，可以被选择性地跳过。为了智能地跳过层而不牺牲准确性，我们发明了一种先验-后验跳过指导机制来决定何时启动层跳过。我们还提出了一种跳过感知的两阶段知识蒸馏算法，以高效地训练标准VLA到DySL-VLA。我们的实验表明，DySL-VLA在Calvin数据集上相比Deer-VLA在成功长度上提高了2.1%，同时将可训练参数减少了85.7倍，并在等精度条件下相对于RoboFlamingo基线提供了3.75倍的速度提升。我们的代码可在https://github.com/PKU-SEC-Lab/DYSL_VLA上获得。
英文摘要：
Vision-Language-Action (VLA) models have shown remarkable success in robotic tasks like manipulation by fusing a language model's reasoning with a vision model's 3D understanding. However, their high computational cost remains a major obstacle for real-world applications that require real-time performance. We observe that the actions within a task have varying levels of importance: critical steps demand high precision, while less important ones can tolerate more variance. Leveraging this insight, we propose DySL-VLA, a novel framework that addresses computational cost by dynamically skipping VLA layers based on each action's importance. DySL-VLA categorizes its layers into two types: informative layers, which are consistently executed, and incremental layers, which can be selectively skipped. To intelligently skip layers without sacrificing accuracy, we invent a prior-post skipping guidance mechanism to determine when to initiate layer-skipping. We also propose a skip-aware two-stage knowledge distillation algorithm to efficiently train a standard VLA into a DySL-VLA. Our experiments indicate that DySL-VLA achieves 2.1% improvement in success length over Deer-VLA on the Calvin dataset, while simultaneously reducing trainable parameters by a factor of 85.7 and providing a 3.75x speedup relative to the RoboFlamingo baseline at iso-accuracy. Our code is available on https://github.com/PKU-SEC-Lab/DYSL_VLA.
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。
