超越路由：表征专家调节与表示在视觉混合专家中的刻画
Beyond Routing: Characterising Expert Tuning and Representation in Vision Mixture-of-Experts
AI总结：
本文研究了视觉混合专家模型中专家调节与表示的特性，通过对比学习训练稀疏门控卷积MoE模型，并利用视觉神经科学工具分析专家的专业化，发现动植物区分主导专家划分，并揭示了专家在更广泛的连续视觉和语义维度上的调节。
摘要：
混合专家（MoE）模型通常通过分析哪些类别被路由到哪些专家来解释。然而，仅靠路由并不能揭示每个专家实际编码的内容。我们训练了稀疏门控卷积MoE模型，并在自然图像上使用对比目标进行训练，利用视觉神经科学工具来表征专家的专业化。从门控级别扩展到专家级别分析，我们测量了每个专家的类别分离度，并利用最吸引人的输入来分析每个专家的调节。从类别级别扩展到特征级别解释，我们通过从人类行为判断数据集（THINGS）中衍生出的语义维度来解释调节。最后，我们使用调节和表征相似性分析来评估在独立初始化下专家分配的稳定性。我们发现，动植物区分主导专家划分，从门控到专家读取都明显，并在独立训练模型中保持稳定。尽管路由统计数据表明相对稀疏的、类别的偏好，但专家分析揭示了更广泛的对连续视觉和语义维度的调节，超出了类别边界。尽管特征调节不同，专家之间表现出相似的类别分离度，这表明超越类别级别分析的解释优势。这些结果表明，视觉MoE中的专家专业化远超类别路由，并通过探测细粒度专家级别调节和表征结构来更好地理解。
英文摘要：
Mixture-of-Experts (MoE) models are often interpreted by analysing which categories are routed to which experts. However, routing alone does not reveal what each expert actually encodes. We train sparsely-gated convolutional MoE models with a contrastive objective on natural images and characterise expert specialisation using tools from visual neuroscience. Extending from gating-level to expert-level analyses, we measure per-expert category separability, and per-expert tuning using the most exciting inputs. Extending from category-level to feature-level explanations, we interpret tuning via semantic dimensions derived from a dataset of human behavioural judgements (THINGS). Finally, we use tuning and representational similarity analysis to assess the stability of expertise-allocation across independent initialisations. We find that an animate-inanimate distinction dominates expert partitioning, apparent from gating through to expert readout, and is stable across independently trained models. Although routing statistics suggest relatively sparse, categorical preferences, expert analyses reveal broader tuning to continuous visual and semantic dimensions that extend beyond category boundaries. Experts exhibit similar category-separability to one another, despite distinct feature tuning, demonstrating the explanatory benefits of moving beyond category-level analyses. Together, these results show that expert specialisation in vision MoEs extends well beyond category routing and is better understood by probing fine-grained expert-level tuning and representational structure.
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。
