SCOPE：视觉感知专家的选择性跨模态编排
Scope: Selective Cross-modal Orchestration of Visual Perception Experts
AI总结：
SCOPE提出编码器混合框架，通过实例级路由为每个图像-文本对动态选择最优视觉编码器，以更少计算量超越堆叠全部编码器的性能。
摘要：
视觉语言模型（VLM）受益于多个视觉编码器，但简单堆叠它们会带来收益递减，同时使推理成本成倍增加。我们提出SCOPE，一种编码器混合（MoEnc）框架，通过实例级路由为每个图像-文本对动态选择一个专用编码器，这与传统MoE中的令牌级路由不同。SCOPE维护一个共享编码器和一个路由编码器池。一个轻量级路由器利用文本提示与共享视觉特征之间的交叉注意力，从路由编码器中选择最优编码器。为训练该路由器，我们引入双熵正则化及辅助损失，以平衡数据集级别的负载分布与实例级别的路由置信度。值得注意的是，采用一个共享加一个路由编码器的SCOPE，其性能优于同时使用全部四个额外编码器的模型，同时计算量减少24%-49%。这表明智能编码器选择优于暴力聚合，挑战了多编码器VLM中的主流范式。
英文摘要：
Vision-language models (VLMs) benefit from multiple vision encoders, but naively stacking them yields diminishing returns while multiplying inference costs. We propose SCOPE, a Mixture-of-Encoders (MoEnc) framework that dynamically selects one specialized encoder per image-text pair via instance-level routing, unlike token-level routing in traditional MoE. SCOPE maintains a shared encoder and a pool of routed encoders. A lightweight router uses cross-attention between text prompts and shared visual features to select the optimal encoder from the routed encoders. To train this router, we introduce dual entropy regularization with auxiliary losses to balance dataset-level load distribution with instance-level routing confidence. Remarkably, SCOPE with one shared plus one routed encoder outperforms models using all four extra encoders simultaneously, while reducing compute by 24-49\%. This demonstrates that intelligent encoder selection beats brute-force aggregation, challenging the prevailing paradigm in multi-encoder VLMs.
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。
