论文HF Daily Papers · arXiv 2611.08373昨天

SLA模型通过持续预训练将3D空间对齐能力注入VLA基座

提出SLA空间-语言-动作模型，基于pi05策略初始化，在持续预训练阶段联合优化动作预测与3D空间表征对齐。引入特征依赖门控机制适配多源异构机器人数据，利用概率状态Dropout增强视觉鲁棒性，且推理架构保持零额外开销。

AI 整理

数据与采集
具身模型

本文由 AI 根据公开内容整理，可能有误，请以原文为准：SLA: Scaling Spatial-Language-Action Model via Continual Pre-training（HF Daily Papers · arXiv 2611.08373）

核心内容

针对现有VLA模型缺乏原生3D空间理解的问题，研究团队提出SLA（Spatial-Language-Action）模型。该方法以预训练pi05策略为初始化基座，将隐式空间对齐扩展至持续预训练阶段并贯穿下游微调。通过构建统一的多模态机器人数据集，SLA实现了跨具身形态、任务与场景的空间知识迁移，解决了传统方法仅在任务适配阶段引入空间监督导致的泛化瓶颈。

技术层面，SLA采用冻结的3D感知模型作为教师端，联合优化动作预测与中间层视觉Embedding的空间对齐。为适配异构数据，设计了特征依赖门控机制自适应聚合多层VLA视觉特征；同时利用部分可用的深度图与相机标定参数辅助RGB观测训练。引入概率状态Dropout迫使模型依赖视觉线索以防过拟合。所有辅助模块仅用于训练，确保推理时可直接复用原VLA权重与架构。

关键看点

持续预训练对齐：基于pi05初始化，在预训练阶段联合优化动作预测与冻结3D感知模型的空间表征对齐，将空间理解内化为基座能力
特征依赖门控：设计自适应门控机制聚合多层VLA视觉Embedding，解决跨具身、跨场景异构数据的空间监督信号适配难题
异构几何标注利用：教师模型融合部分可用深度图、相机标定及RGB观测，有效利用非标准化几何标注提升3D表征学习效率
推理零开销：概率状态Dropout等辅助组件仅限训练期使用，推理阶段完全复用原始VLA权重与架构，无额外计算负载

工程与产业观察

SLA验证了将3D空间对齐前置到VLA持续预训练的有效性，为具身智能提供了从2D语义向3D物理世界对齐的工程范式。其特征依赖门控与异构几何标注利用策略，对解决多源机器人数据混合训练中的模态不对齐问题具有直接参考价值，且推理零开销设计保障了产业部署可行性。

英文摘要 (Abstract)

Reliable robotic manipulation requires both semantic understanding and precise control in three-dimensional space. However, vision-language-action (VLA) foundations largely inherit visual representations from 2D vision-language pre-training, while spatial supervision is often introduced only during task-specific adaptation. Developing spatial understanding as a transferable foundation-model capability therefore calls for learning from broader robot experience. We present SLA, a Spatial-Language-Action Model that scales implicit spatial alignment to continual pre-training and maintains it during downstream post-training. Initialized from a pretrained pi05 policy, SLA jointly optimizes action prediction and alignment between intermediate visual embeddings and spatial representations from a frozen 3D perception model. We curate and unify heterogeneous robot datasets to support spatial learning across diverse embodiments, tasks, and scenes. To accommodate this diversity, we introduce feature-dependent gates that adaptively aggregate visual embeddings from multiple VLA layers for spatial supervision. The teacher further exploits partially available depth and camera calibration alongside RGB observations, enabling effective use of heterogeneous geometric annotations. Probabilistic state dropout encourages reliance on visual cues and mitigates overfitting to demonstration trajectories. All auxiliary components are used only during training, allowing direct reuse of pretrained VLA weights and preserving the original inference architecture.

Hugging Face Daily Papers 热度：24 upvotes 作者：Fuhao Li, Wenxuan Song, Qizhou Huang, Ziyang Zhou, Jingbo Wang, Junwu E, Pengwei Wang, Yan Wang, Shunbo Zhou, Haoang Li