CleanVideo：面向文本到视频扩散模型的自适应概念擦除
CleanVideo: Adaptive Concept Erasure for Text-to-Video Diffusion Models
- Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)（穆罕默德·本·扎耶德人工智能大学）
- University of Electronic Science and Technology of China (UESTC)（电子科技大学）
- The Hong Kong University of Science and Technology (HKUST)（香港科技大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
CleanVideo提出一种基于三模态门控的低维子空间干预框架，实现视频扩散模型中的自适应概念擦除，在保持视觉保真度与时间一致性的同时有效移除目标概念。
AI中文摘要：
概念擦除旨在从预训练生成模型中选择性地消除不期望的视觉语义，同时不损害其通用效用。将概念擦除从图像扩展到视频并非易事。目标概念会逐渐出现，并在不同帧和去噪步骤中发生变化。因此，固定的干预措施可能错过目标或引入模糊、抖动和内容失真。我们提出CleanVideo，一种选择性擦除框架，通过三模态门控机制执行低维子空间干预。通过联合处理时空视觉特征、时间步信号和文本语义，CleanVideo确定在何处、何时以及是否进行干预，在可明确定义替代概念时将擦除内容引导至自然的替代概念，同时保留非目标内容。在三个视频扩散模型上的实验表明，CleanVideo能有效擦除目标概念，同时保持视觉保真度和时间一致性，在帧级和视频级评估以及保护流程保持完整时的概念恢复攻击下均优于现有基线。
英文摘要：
Concept erasure aims to selectively eliminate undesired visual semantics from pre-trained generative models without compromising their general utility. Extending concept erasure from images to video is nontrivial. Target concepts emerge gradually and vary across frames and denoising steps. As a result, fixed interventions may miss the target or introduce blurring, jitter, and content distortion. We propose CleanVideo, a selective erasure framework that performs low-dimensional subspace intervention controlled by a tri-modal gating mechanism. By jointly processing spatiotemporal visual features, timestep signals, and textual semantics, CleanVideo determines where, when, and whether to intervene, steering erased content toward natural surrogate concepts when such surrogates can be clearly defined while preserving non-target content. Experiments on three video diffusion models show that CleanVideo effectively erases target concepts while maintaining visual fidelity and temporal coherence, outperforming existing baselines under frame-level and video-level evaluations and under concept-recovery attacks when the protected pipeline remains intact.
