恢复VLA模型中的语言基础：无需训练的注意力重校准方法
Restoring Linguistic Grounding in VLA Models via Train-Free Attention Recalibration
AI总结：
针对VLA模型在分布外指令下出现“语言盲视”问题，提出无需训练的注意力重校准方法IGAR，通过调整注意力分布恢复语言指令影响，在LIBERO基准和真实机器人上有效减少错误执行。
摘要：
视觉-语言-动作（VLA）模型使机器人能够直接根据自然语言指令执行操作任务，并日益被视为通用机器人策略的基础。然而，它们在分布外（OOD）指令下的可靠性尚未得到充分探索。在本文中，我们揭示了一个关键的失败模式：即使语言指令与场景矛盾，VLA策略仍继续执行视觉上合理的动作。我们将这种现象称为语言盲视，即VLA策略在动作生成过程中优先考虑视觉先验而非指令语义。为了系统分析这一问题，我们引入了ICBench，一个基于LIBERO数据集构建的诊断基准，通过注入受控的OOD指令矛盾（同时保持视觉环境不变）来探测语言-动作耦合。对三种代表性VLA架构（包括Pi0、Pi0.5和OpenVLA OFT）的评估表明，尽管指令逻辑上不可能，这些模型仍频繁成功完成任务，揭示了动作生成中的强烈视觉偏差。为了缓解这一问题，我们提出了指令引导的注意力重校准（IGAR），一种无需训练的推理时机制，通过重新平衡注意力分布来恢复语言指令的影响。IGAR无需重新训练或架构修改，可直接应用于现有VLA模型。在30个LIBERO任务上的实验表明，IGAR在OOD矛盾指令下显著减少了错误执行，同时保持了基线任务性能。我们还在真实的Franka机器人臂上验证了该方法，IGAR有效防止了由不一致指令触发的操作。
英文摘要：
Vision-Language-Action (VLA) models enable robots to perform manipulation tasks directly from natural language instructions and are increasingly viewed as a foundation for generalist robotic policies. However, their reliability under Out-of-Distribution (OOD) instructions remains underexplored. In this paper, we reveal a critical failure mode in which VLA policies continue executing visually plausible actions even when the language instruction contradicts the scene. We refer to this phenomenon as linguistic blindness, where VLA policies prioritize visual priors over instruction semantics during action generation. To systematically analyze this issue, we introduce ICBench, a diagnostic benchmark constructed from the LIBERO dataset that probes language-action coupling by injecting controlled OOD instruction contradictions while keeping the visual environment unchanged. Evaluations on three representative VLA architectures, including Pi0, Pi0.5 and OpenVLA OFT, show that these models frequently succeed at tasks despite logically impossible instructions, revealing a strong visual bias in action generation. To mitigate this issue, we propose Instruction-Guided Attention Recalibration (IGAR), a train-free inference-time mechanism that rebalances attention distributions to restore the influence of language instructions. IGAR operates without retraining or architectural modification and can be directly applied to existing VLA models. Experiments across 30 LIBERO tasks demonstrate that IGAR substantially reduces erroneous execution under OOD contradictory instructions while preserving baseline task performance. We additionally validate the approach on a real Franka robotic arm, where IGAR effectively prevents manipulation triggered by inconsistent instructions.
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。
