TAG-MoE：面向统一生成式混合专家的任务感知门控
TAG-MoE: Task-Aware Gating for Unified Generative Mixture-of-Experts
把任务语义（范围、类型、保留程度）注入 MoE 门控网络，用预测对齐正则化让路由决策与高层任务意图对齐，可直接迁移到任何需要多任务共享参数的扩散 Transformer 中，缓解局部编辑与主体驱动生成之间的任务干扰。
原文依据 we propose a novel framework to inject semantic intent into MoE routing
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论统一的图像生成与编辑模型在稠密扩散 Transformer 架构中会遭受严重的任务干扰，共享的参数空间必须在相互冲突的目标（例如局部编辑与主体驱动生成）之间做出折中。稀疏混合专家（MoE）范式是一种有前景的解决方案，但其门控网络仍然是任务无关的，仅依据局部特征运行，无法感知全局任务意图。这种任务无关性阻碍了有意义的专门化，也无法解决底层的任务干扰。本文提出一种新框架，将语义意图注入 MoE 路由。我们引入分层任务语义标注方案，以构建结构化的任务描述符（例如范围、类型、保留程度）。随后我们设计预测对齐正则化，使内部路由决策与任务的高层语义对齐。该正则化使门控网络从任务无关的执行器演化为调度中心。我们的模型有效缓解了任务干扰，在保真度和质量上优于稠密基线，并且我们的分析表明专家会自然形成清晰且语义相关的专门化。
英文原文摘要 1214 字符
Unified image generation and editing models suffer from severe task interference in dense diffusion transformers architectures, where a shared parameter space must compromise between conflicting objectives (e.g., local editing v.s. subject-driven generation). While the sparse Mixture-of-Experts (MoE) paradigm is a promising solution, its gating networks remain task-agnostic, operating based on local features, unaware of global task intent. This task-agnostic nature prevents meaningful specialization and fails to resolve the underlying task interference.In this paper, we propose a novel framework to inject semantic intent into MoE routing. We introduce a Hierarchical Task Semantic Annotation scheme to create structured task descriptors (e.g., scope, type, preservation). We then design Predictive Alignment Regularization to align internal routing decisions with the task's high-level semantics. This regularization evolves the gating network from a task-agnostic executor to a dispatch center. Our model effectively mitigates task interference, outperforming dense baselines in fidelity and quality, and our analysis shows that experts naturally develop clear and semantically correlated specializations.