雕塑记忆：基于动态掩码与概念感知优化的扩散模型多概念遗忘
Sculpting Memory: Multi-Concept Forgetting in Diffusion Models via Dynamic Mask and Concept-Aware Optimization
AI总结：
针对扩散模型多概念遗忘中的不稳定与质量下降问题，提出结合动态掩码与概念感知损失的新型遗忘框架，通过自适应梯度掩码和知识蒸馏正则化实现稳定高保真的多概念遗忘。
摘要：
文本到图像（T2I）扩散模型在根据文本提示生成高质量图像方面取得了显著成功。然而，其存储海量知识的能力在需要选择性遗忘的场景中引发了担忧，例如移除受版权保护的内容、减少偏见或消除有害概念。尽管现有的机器遗忘方法能够移除某些概念，但由于不稳定性、残留知识持续存在以及生成质量下降，它们在多概念遗忘方面面临困难。为了应对这些挑战，我们提出了Dynamic Mask与Concept-Aware Loss，这是一种专为扩散模型多概念遗忘设计的新型遗忘框架。我们的Dynamic Mask机制根据当前优化状态自适应地更新梯度掩码，允许进行选择性权重修改，从而防止干扰无关知识。此外，我们的Concept-Aware Loss通过在父类对齐上强制执行语义一致性来显式指导遗忘过程，同时基于知识蒸馏的正则化损失确保在顺序遗忘期间先前已遗忘的概念保持被遗忘状态。我们进行了大量实验来评估我们的方法。结果表明，我们的方法在遗忘有效性、输出保真度和语义连贯性方面优于现有的遗忘技术，特别是在多概念场景中。我们的工作为生成模型中稳定且高保真的遗忘提供了一个有原则且灵活的框架。代码将公开发布。
英文摘要：
Text-to-image (T2I) diffusion models have achieved remarkable success in generating high-quality images from textual prompts. However, their ability to store vast amounts of knowledge raises concerns in scenarios where selective forgetting is necessary, such as removing copyrighted content, reducing biases, or eliminating harmful concepts. While existing unlearning methods can remove certain concepts, they struggle with multi-concept forgetting due to instability, residual knowledge persistence, and generation quality degradation. To address these challenges, we propose \textbf{Dynamic Mask coupled with Concept-Aware Loss}, a novel unlearning framework designed for multi-concept forgetting in diffusion models. Our \textbf{Dynamic Mask} mechanism adaptively updates gradient masks based on current optimization states, allowing selective weight modifications that prevent interference with unrelated knowledge. Additionally, our \textbf{Concept-Aware Loss} explicitly guides the unlearning process by enforcing semantic consistency through superclass alignment, while a regularization loss based on knowledge distillation ensures that previously unlearned concepts remain forgotten during sequential unlearning. We conduct extensive experiments to evaluate our approach. Results demonstrate that our method outperforms existing unlearning techniques in forgetting effectiveness, output fidelity, and semantic coherence, particularly in multi-concept scenarios. Our work provides a principled and flexible framework for stable and high-fidelity unlearning in generative models. The code will be released publicly.
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。
