当安全相互冲突：通过自适应安全引导消解文本到图像扩散中的多类别有害冲突
When Safety Collides: Resolving Multi-Category Harmful Conflicts in Text-to-Image Diffusion via Adaptive Safety Guidance
对做扩散模型安全对齐的人，这篇给出了一个推理时、免训练的多类别安全方向动态选择机制：不再用多类关键词平均出的单一有害区域，而是按当前生成状态识别最对齐的有害类别并只沿该方向引导，可直接插到潜空间或文本空间的安全模块上。
原文依据 we propose Conflict-aware Adaptive Safety Guidance (CASG), a training-free framework that dynamically identifies and applies the category-aligned safety direction during generation
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文本到图像（T2I）扩散模型在生成高质量图像方面取得了显著进展，同时也引发了关于有害内容生成的潜在安全担忧。已有研究提出了基于安全引导的方法，通过将生成过程引导远离有害区域来减少有害输出，其中这些区域是基于预定义关键词在多个有害类别上平均得到的。然而，这类方法未能刻画不同有害类别之间复杂的相互作用，导致出现「有害冲突」：缓解某一类有害内容可能会无意中放大另一类，从而抬高整体有害率。为解决该问题，我们提出冲突感知自适应安全引导（CASG），这是一个免训练框架，能够在生成过程中动态识别并施加与类别对齐的安全方向。CASG 由两个部分组成：（i）冲突感知类别识别（CaCI），用于识别与模型不断演化的生成状态最为对齐的有害类别；（ii）冲突消解引导施加（CrGA），仅沿所识别出的类别施加安全引导，以避免多类别之间的干扰。CASG 可同时应用于潜空间与文本空间的安全防护。在 T2I 安全基准上的实验表明，CASG 取得了最优性能，与现有方法相比将有害率最多降低 15.4%。
英文原文摘要 1420 字符
Text-to-Image (T2I) diffusion models have demonstrated significant advancements in generating high-quality images, while raising potential safety concerns regarding harmful content generation. Safety-guidance-based methods have been proposed to mitigate harmful outputs by steering generation away from harmful zones, where the zones are averaged across multiple harmful categories based on predefined keywords. However, these approaches fail to capture the complex interplay among different harm categories, leading to "harmful conflicts" where mitigating one type of harm may inadvertently amplify another, thus increasing overall harmful rate. To address this issue, we propose Conflict-aware Adaptive Safety Guidance (CASG), a training-free framework that dynamically identifies and applies the category-aligned safety direction during generation. CASG is composed of two components: (i) Conflict-aware Category Identification (CaCI), which identifies the harmful category most aligned with the model's evolving generative state, and (ii) Conflict-resolving Guidance Application (CrGA), which applies safety steering solely along the identified category to avoid multi-category interference. CASG can be applied to both latent-space and text-space safeguards. Experiments on T2I safety benchmarks demonstrate CASG's state-of-the-art performance, reducing the harmful rate by up to 15.4% compared to existing methods.