Agentic-VLA：视觉-语言-动作模型的高效在线自适应
Agentic-VLA: Efficient Online Adaptation for Vision-Language-Action Models
AI总结：
提出Agentic-VLA框架，通过自适应奖励合成、语言引导探索和经验记忆三大创新，实现VLA模型在线高效自适应，在LIBERO基准上长时域任务提升12.3%，单样本学习提升28.5%，并实现零演示跨任务迁移。
摘要：
视觉-语言-动作（VLA）模型通过利用预训练的视觉-语言表示，已成为机器人操作领域的一种有前景的范式。然而，当前的VLA训练方法存在两个关键局限性：对新环境的泛化能力差，以及需要大量演示数据导致的训练效率低下。我们提出Agentic-VLA，一种智能训练框架，通过三项关键创新使VLA能够在线高效自适应：（1）自适应奖励合成，根据VLA当前能力和任务复杂度动态生成并调整奖励函数，将复杂任务分解为可学习的子目标以进行课程学习；（2）语言引导探索，其中评论模型提供结构化指导以实现系统化探索，而非随机采样；（3）经验记忆，存储和检索与任务相关的策略权重，用于相似任务的预热启动自适应。我们在LIBERO基准上评估Agentic-VLA，取得了显著改进：长时域任务提升12.3%，单样本学习提升28.5%，并在无需任务特定演示的情况下实现从0%到31.2%的跨任务迁移。与现有在线自适应方法相比，我们的框架还实现了2.4倍的收敛速度提升。除LIBERO外，Agentic-VLA在双臂RoboTwin 2.0基准（包括其随机困难设置）上仍保持优势。这些结果使Agentic-VLA成为迈向真正自适应、可在部署中持续学习的VLA系统的重要一步。
英文摘要：
Vision-Language-Action (VLA) models have emerged as a promising paradigm for robotic manipulation by leveraging pre-trained vision-language representations. However, current VLA training methods suffer from two critical limitations: poor generalization to novel environments and low training efficiency requiring extensive demonstrations. We introduce Agentic-VLA, an agentic training framework that enables VLAs to efficiently adapt online through three key innovations: (1) Adaptive Reward Synthesis, which dynamically generates and adjusts reward functions based on the VLA's current capabilities and task complexity, decomposing complex tasks into learnable sub-goals for curriculum learning; (2) Language-Guided Exploration, where a critic model provides structured guidance for systematic exploration rather than random sampling; and (3) Experience Memory,which stores and retrieves task-relevant policy weights for warm-starting adaptation to similar tasks. We evaluate Agentic-VLA on the LIBERO benchmark, achieving substantial improvements: +12.3% on long-horizon tasks, +28.5% in 1-shot learning, and enabling cross-task transfer from 0% to 31.2% without task-specific demonstrations. Our framework also demonstrates 2.4x faster convergence compared to existing online adaptation methods. Beyond LIBERO, Agentic-VLA retains its advantage on the dual-arm RoboTwin 2.0 benchmark, including under its randomized Hard setting. These results establish Agentic-VLA as a significant step toward truly adaptive VLA systems capable of continuous learning in deployment.
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。
