AI 译本 · Equalang
新上架PDF · 2026年11月翻译
加载中…
加载中…
arXiv · 2026
超越数据规模扩展：以表征为中心的视觉-语言-动作模型持续预训练
Senqiao Yang, Chengyao Wang, Yuxin Chen, Zixuan Wang et al.
Scaling robot data is crucial for building generalist Vision-Language-Action (VLA) models, yet robot trajectories are harder to scale than web-scale image-text data because embodied collection is costly and sparsely covers the physical world. This makes representation quality a central bottleneck: under a fixed robot-data budget, continued pre-training must turn limited trajectories into transferable visual-action knowledge rather than merely fit actions. We propose VLAct, a VLA-oriented VLM backbone trained on broad, heterogeneous, multi-embodiment robot data before task-specific fine-tuning. VLAct preserves the broad VLM prior and encourages shared action semantics across embodiments through VLM-prior preservation, multi-head continuous action co-supervision, and a partially unified cross-embodiment action layout, while allowing task-specific action heads during fine-tuning. Across simulation, real-world, and unseen-embodiment transfer, VLAct consistently improves downstream performance under fixed fine-tuning protocols. On LIBERO-Plus and RoboTwin 2.0, VLAct surpasses industrial VLA systems including ABot-M0 and LingBot-VLA, achieving success rates of 82.6% and 92.5%. On RoboDojo, VLAct ranks sixth among all policies by success rate and outperforms all explicitly designated world-action model (WAM) entries on both metrics. Most notably, on RoboCasa-GR1, an unseen humanoid embodiment, VLAct using only 20% of downstream trajectories outperforms the full-data GR00T-N1.6 baseline. These results are obtained using fully open-source data and only a 16-GPU training setup, showing that representation-centric continued pre-training can deliver highly competitive performance under a modest compute budget and is an important independent axis of VLA progress beyond data scaling.
扩展机器人数据规模对于构建通用视觉-语言-动作（VLA）模型至关重要。然而，与网络规模的图文数据相比，机器人轨迹更难扩展，因为具身数据采集成本高昂，且对物理世界的覆盖较为稀疏。因此，表征质量成为关键瓶颈：在机器人数据预算固定的情况下，持续预训练必须将有限的轨迹转化为可迁移的视觉-动作知识，而不能仅仅拟合动作。我们提出了 VLAct，这是一种面向 VLA 的视觉语言模型（VLM）骨干网络，先使用广泛、异质且涵盖多种具身形态的机器人数据进行训练，再针对特定任务微调。VLAct 通过保留 VLM 先验、多头连续动作协同监督，以及部分统一的跨具身动作布局，在保留 VLM 广泛先验的同时，促进不同具身形态之间共享动作语义；此外，它还允许在微调阶段使用特定任务的动作头。在仿真、真实世界以及向未见具身形态迁移的测试中，在固定微调方案下，VLAct 均能稳定提升下游性能。在 LIBERO-Plus 和 RoboTwin 2.0 上，VLAct 超越了 ABot-M0 和 LingBot-VLA 等工业级 VLA 系统，成功率分别达到 82.6% 和 92.5%。在 RoboDojo 上，按成功率排名，VLAct 在所有策略中位列第六，并且在两项指标上均优于所有明确标注为世界-动作模型（WAM）的参赛方案。尤为值得一提的是，在采用此前未见过的人形机器人具身形态的 RoboCasa-GR1 上，VLAct 仅使用 20% 的下游轨迹数据，表现便超过了使用全部数据训练的 GR00T-N1.6 基线。这些结果基于完全开源的数据，仅使用 16 张 GPU 进行训练，表明以表征为中心的持续预训练能够在适度的算力预算下实现极具竞争力的性能，并且是数据规模扩展之外推动 VLA 进步的一条重要且独立的路径。
PDF · 2026年11月翻译