发表机构
Peking University; National University of Singapore; NVIDIA; Impossible Research（北京大学; 新加坡国立大学; 英伟达; Impossible Research）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对VLM机器人智能体令牌开销大的问题，提出PyRUA-Lean框架，通过选择性观测和原语组合，在相等调用预算下将成功率从63.1%提至71.7%，并减少65%令牌。
AI 中文摘要
视觉语言模型（VLM）智能体可通过视觉反馈和动作原语控制机器人，但重复的模型调用和冗余的观测会产生大量令牌开销。我们提出PyRUA-Lean，一个交互式代码执行框架，将反馈驱动的原语组合与选择性观测相结合：智能体将经典机器人原语和学习到的视觉-语言-动作（VLA）策略组合成Python单元，这些单元执行条件检查与局部重试，仅返回明确请求的图像和用于重新规划的状态反馈。在来自LIBERO-PRO、RoboTwin 2.0和RoboCasa365的700个模拟任务实例中，我们将PyRUA-Lean与使用相同GPT-6 Astra规划器和底层机器人原语的工具调用基线进行比较。在相等的LLM调用预算下，PyRUA-Lean将整体成功率从63.1%提升至71.7%。在两个智能体均解决的实例中，它使用的LLM调用次数减少49%，输入令牌减少65%。
英文摘要
Vision language model (VLM) agents can control robots through visual feedback and action primitives, but repeated model invocations and redundant observations incur substantial token overhead. We introduce PyRUA-Lean, an interactive code-execution framework that couples feedback-driven primitive composition with selective observation: the agent composes classical robot primitives and learned vision-language-action (VLA) policies into Python cells that perform conditional checks and local retries, returning only explicitly requested images and state feedback for replanning. Across 700 simulated task instances from LIBERO-PRO, RoboTwin 2.0, and RoboCasa365, we compare PyRUA-Lean with a tool-calling baseline using the same GPT-6 Astra planner and underlying robot primitives. Under equal LLM-call budgets, PyRUA-Lean increases overall success from 63.1% to 71.7%. On instances solved by both agents, it uses 49% fewer LLM calls and 65% fewer input tokens.
