BridgeWAM：通过潜在桥接查询桥接视频与动作专家
BridgeWAM: Bridging Video and Action Experts via Latent Bridge Queries
AI总结：
本文提出BridgeWAM，通过32个潜在桥接查询聚合多模态先验，解耦视频与动作专家深度，在LIBERO及LIBERO-Plus数据集上取得高任务成功率，同时大幅降低推理延迟与运算量。
摘要：
世界动作模型（World Action Models, WAMs）通常通过密集的分层混合Transformer（Mixture-of-Transformers, MoT）注意力机制耦合视频专家与动作专家，这使得动作预测易受无关视觉特征干扰，且要求专家深度匹配，阻碍了选择性视觉特征的使用并引入冗余计算。本文提出专家桥接（Bridge-of-Experts, BoE），实例化为BridgeWAM，其通过32个潜在桥接查询（Latent Bridge Queries, LBQs）将视觉、语言和本体感受先验聚合到共享的有限潜在空间中。联合视频-动作监督使动作梯度引导视频专家生成更适合动作预测的表示，同时动作专家通过交叉注意力学习选择性检索任务相关信息。解耦专家深度并在动作去噪过程中复用LBQs，得到了一个50.6亿参数的模型，其视频专家为30层，动作专家为2层，在LIBERO数据集上达到98.25%的成功率，在LIBERO-Plus数据集上达到72.24%的汇总成功率，证明了其在超轻量动作头的情况下具备有效的任务执行能力和强泛化性。消融实验结果支持联合表示适配和统一LBQ条件设置。与Fast-WAM相比，BridgeWAM在仿真分析中将单样本推理浮点运算量（FLOPs）降低了28.72%，并将真实机器人延迟降至77.1毫秒，实现了2.47倍的加速。定性实验结果进一步表明，该模型在存在扰动时会优先关注目标物体和任务目的地。
英文摘要：
World Action Models (WAMs) commonly couple video and action experts through dense, layer-wise Mixture-of-Transformers (MoT) attention, exposing action prediction to irrelevant visual features while enforcing matched expert depths. This coupling hinders selective visual-feature use and introduces redundant computation. We introduce Bridge-of-Experts (BoE), instantiated as BridgeWAM, which aggregates visual, language, and proprioceptive priors through 32 Latent Bridge Queries (LBQs) into a shared, limited latent space. Joint video-action supervision lets action gradients guide the Video Expert toward representations better suited for action prediction, while the Action Expert learns to selectively retrieve task-relevant information through cross-attention. Decoupling expert depths and reusing LBQs during action denoising yield a 5.06B model with a 30-layer Video Expert and a two-layer Action Expert, achieving 98.25% success on LIBERO and 72.24% pooled success on LIBERO-Plus, demonstrating effective task execution and strong generalization with an ultralight action head. Ablations support joint representation adaptation and unified LBQ conditioning. Relative to Fast-WAM, BridgeWAM reduces single-sample inference FLOPs by 28.72% in simulation profiling and real-robot latency to 77.1 ms (a 2.47-fold speedup). Qualitative experimental results further indicate preferential attention to target objects and task destinations across perturbations.
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。
