ActionMap: 基于体素动作热图的机器人策略学习
ActionMap: Robot Policy Learning via Voxel Action Heatmap
AI总结：
提出ActionMap，一种将动作空间建模为体素热图的动作解码器，替代现有VLA模型中的单点预测器，在LIBERO仿真和真实Franka操作中提升性能和数据效率。
摘要：
视觉-语言-动作（VLA）模型在骨干网络、训练方法和数据规模方面快速发展，但将骨干网络隐藏状态转换为连续控制信号的动作解码器几乎没有变化，在大多数现有VLA中仍然是单点预测器。无论是通过自回归词元箱、L1回归还是流匹配去噪实现，所得解码器都将动作空间视为无结构的，在训练期间未利用相邻动作的几何邻近性。为了改进这一点，我们引入了ActionMap，一种体素热图动作头，可以插入现有VLA中替换其原生动作解码器。对于每个新动作，该头预测动作空间上的体素热图，其中每个体素直接存储对应动作的概率。在LIBERO仿真和真实Franka操作中，我们的热图头在匹配训练步数下超越了两种架构不同的骨干网络（例如，在LIBERO四套件平均上比OpenVLA-OFT的L1回归头高出8.2%），在两种骨干网络上以相当或更快的速度收敛，并且在低训练数据下保持显著更高的数据效率。跨骨干网络的一致性表明，动作表示是VLA性能的一个真正杠杆，与进一步的骨干网络或方法缩放不同。项目页面：此 https URL。
英文摘要：
Vision-language-action (VLA) models have advanced rapidly across backbones, training recipes, and data scale, yet the action decoder, which converts the backbone's hidden state into a continuous control signal, has barely changed and remains a single-point predictor across the majority of current VLAs. Whether implemented via autoregressive token bins, L1 regression, or flow-matching denoising, the resulting decoder treats the action space as unstructured, leaving the geometric proximity of neighboring actions unexploited during training. To advance this, we introduce ActionMap, a voxel heatmap action head that drops into an existing VLA in place of its native action decoder. For each new action, the head predicts a voxel heatmap over the action space, where each voxel directly stores the probability of the corresponding action. Across LIBERO simulation and real-world Franka manipulation, our heatmap head surpasses two architecturally distinct backbones at matched training steps (e.g., +8.2% over OpenVLA-OFT's L1 regression head on the LIBERO four-suite average), converges at comparable or faster rates on both backbones, and remains markedly more data-efficient at low training data. The cross-backbone consistency indicates that action representation is a real lever for VLA performance, distinct from further backbone or recipe scaling. Project Page: https://showlab.github.io/ActionMap/.
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。
