重新思考视觉-语言-动作模型的扩展：对齐、混合与正则化
Rethinking Visual-Language-Action Model Scaling: Alignment, Mixture, and Regularization
- Renmin University of China（中国人民大学）
- BeingBeyond
- Peking University（北京大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本研究重新审视视觉-语言-动作模型的扩展问题，探讨了对齐、混合与正则化在机器人控制中的关键作用，挑战了传统假设并提供实践指导。
AI中文摘要：
尽管视觉-语言-动作（VLA）模型在通用机器人控制中展现出强大潜力，但仍然不清楚标准的
英文摘要：
While Vision-Language-Action (VLA) models show strong promise for generalist robot control, it remains unclear whether -- and under what conditions -- the standard "scale data" recipe translates to robotics, where training data is inherently heterogeneous across embodiments, sensors, and action spaces. We present a systematic, controlled study of VLA scaling that revisits core training choices for pretraining across diverse robots. Using a representative VLA framework that combines a vision-language backbone with flow-matching, we ablate key design decisions under matched conditions and evaluate in extensive simulation and real-robot experiments. To improve the reliability of real-world results, we introduce a Grouped Blind Ensemble protocol that blinds operators to model identity and separates policy execution from outcome judgment, reducing experimenter bias. Our analysis targets three dimensions of VLA scaling. (1) Physical alignment: we show that a unified end-effector (EEF)-relative action representation is critical for robust cross-embodiment transfer. (2) Embodiment mixture: we find that naively pooling heterogeneous robot datasets often induces negative transfer rather than gains, underscoring the fragility of indiscriminate data scaling. (3) Training regularization: we observe that intuitive strategies, such as sensory dropout and multi-stage fine-tuning, do not consistently improve performance at scale. Together, this study challenge some common assumptions about embodied scaling and provide practical guidance for training large-scale VLA policies from diverse robotic data. Project website: https://research.beingbeyond.com/rethink_vla
