用对抗性锚点让扩散模型删掉不良概念，同时保留其他内容
AdvAnchor: Enhancing Diffusion Model Unlearning with Adversarial Anchors
- 设计对抗性文本锚点，模拟不良概念但剔除关键特征
- 在删除不良概念时，保持模型整体性能更优
- 适合需要安全可控生成的AI应用开发者
文本到图像扩散模型的安全问题促使研究者通过微调实现不良概念的消除。现有方法通常将不安全提示的预测分布对齐到预定义文本锚点，但存在消除不良概念与保留其他概念之间的显著性能权衡。本文系统分析了不同文本锚点对去学习效果的影响，据此提出新方法AdvAnchor：生成逼近不良概念嵌入的对抗性锚点，同时有选择地排除其核心属性，以实现有效擦除并维持模型整体性能。大量实验表明，AdvAnchor优于当前最先进方法。代码已公开于 https://anonymous.4open.science/r/AdvAnchor。
原文摘要 · Abstract (English)
Security concerns surrounding text-to-image diffusion models have driven researchers to unlearn inappropriate concepts through fine-tuning. Recent fine-tuning methods typically align the prediction distributions of unsafe prompts with those of predefined text anchors. However, these techniques exhibit a considerable performance trade-off between eliminating undesirable concepts and preserving other concepts. In this paper, we systematically analyze the impact of diverse text anchors on unlearning performance. Guided by this analysis, we propose AdvAnchor, a novel approach that generates adversarial anchors to alleviate the trade-off issue. These adversarial anchors are crafted to closely resemble the embeddings of undesirable concepts to maintain overall model performance, while selectively excluding defining attributes of these concepts for effective erasure. Extensive experiments demonstrate that AdvAnchor outperforms state-of-the-art methods. Our code is publicly available at https://anonymous.4open.science/r/AdvAnchor.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。