MapRoute：通过语义路由实现精确概念擦除的映射器
MapRoute:Precise-Concept Erasing Mappers via Semantic Routing
在冻结的文本编码器之后插入轻量线性映射模块（Mappers），依据文本嵌入与目标概念嵌入的余弦相似度动态激活 top-K 模块，为扩散模型的概念擦除提供了一种免全模型微调、可插拔的干预手段，可直接复用于个性化生成中的敏感概念抑制与版权合规。
原文依据 dynamically activates the top-K Mappers most relevant to the input prompt, based on cosine similarity between the text embedding
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论当代商业与开源扩散模型在文本到图像生成方面表现卓越，推动了创意设计与内容创作中的广泛应用。然而，版权保护、隐私合规或个性化定制等正当需求，往往要求从预训练模型中移除特定的语义概念。现有的概念擦除方法存在两个关键局限：（1）抑制不彻底，模型仍会偶尔生成包含目标概念的图像；（2）语义选择性差，会降低无关概念的生成质量并损害模型的整体效用。为应对这些挑战，我们提出 MapRoute，一个基于动态路由的轻量级、语义感知的概念擦除框架。该方法引入一组模块化组件——称为映射器（Mapper）——置于冻结的预训练文本编码器之后。每个映射器学习从目标概念到替代概念的线性映射。推理时，系统根据文本嵌入与所有目标概念嵌入之间的余弦相似度，动态激活与输入提示最相关的前 K 个映射器，并依次施加其变换。这种由输入驱动、模块化的干预方式实现了精确、按需的擦除，同时避免对无关语义造成不必要的干扰。大量实验表明，MapRoute 能有效抑制指定概念，同时显著减少对无关概念的附带损害。由于无需对全模型进行微调，本方法完全避免了参数漂移与概念侵蚀。此外，MapRoute 在生成保真度、语义一致性以及多概念擦除场景的可扩展性方面均优于最优基线方法。
英文原文摘要 1888 字符
Contemporary commercial and open-source diffusion models have demonstrated remarkable performance in text-to-image generation, enabling widespread applications in creative design and content creation. However, legitimate requirements--such as copyright protection, privacy compliance, or personalized customization--often necessitate the removal of specific semantic concepts from pretrained models. Existing concept erasure methods suffer from two critical limitations: (1) **Incomplete suppression**, where the model still occasionally generates images containing the target concept; (2) **Poor semantic selectivity**, which degrades the generation quality of unrelated concepts and compromises overall model utility.To address these challenges, we propose **`MapRoute`**, a lightweight, semantics-aware concept erasure framework based on dynamic routing. Our approach introduces a set of modular components--termed *Mappers*--placed after a frozen pretrained text encoder. Each Mapper learns a linear mapping from a target concept to a surrogate concept. During inference, the system dynamically activates the top-K Mappers most relevant to the input prompt, based on cosine similarity between the text embedding and all the target concept embeddings, and applies their transformations sequentially. This input-driven, modular intervention enables precise, on-demand erasure while avoiding unnecessary interference with irrelevant semantics.Extensive experiments demonstrate that **`MapRoute`** effectively suppresses specified concepts while significantly reducing collateral damage to unrelated concept. By operating without full-model fine-tuning, our method entirely avoids parameter drift and concept erosion. Moreover, **`MapRoute`** outperforms state-of-the-art baselines in terms of generation fidelity, semantic consistency, and scalability to multi-concept erasure scenarios.