封堵后门：利用知识蒸馏消除扩散模型中的对抗性文本触发器
Sealing The Backdoor: Unlearning Adversarial Text Triggers In Diffusion Models Using Knowledge Distillation
AI总结：
针对文本到图像扩散模型易受文本触发器后门攻击的问题，提出SKD-CAG方法，通过带交叉注意力引导的自知识蒸馏消除触发器与中毒输出的关联，在保持生成质量的同时实现100%像素后门和93%风格攻击消除准确率。
摘要：
文本到图像扩散模型彻底改变了生成式人工智能，但它们对后门攻击的脆弱性构成了重大安全风险。攻击者可以在训练数据中注入难以察觉的文本触发器，导致模型生成被操纵的输出。尽管分类模型中基于文本的后门防御已得到充分探索，但生成模型缺乏有效的缓解技术。我们通过选择性地消除模型学习到的对抗性文本触发器与中毒输出之间的关联，同时保留整体生成质量来解决这个问题。我们的方法——带有交叉注意力引导的自知识蒸馏（SKD-CAG）——利用知识蒸馏引导模型纠正对中毒提示的响应，同时利用后门模型在没有触发器时仍能生成干净输出的事实来保持图像质量。通过交叉注意力机制，SKD-CAG在注意力层面中和后门影响，确保针对性地消除对抗性效果。大量实验表明，我们的方法优于现有方法，在不牺牲鲁棒性或图像保真度的情况下，对像素后门的消除准确率达到100%，对基于风格的攻击达到93%。我们的研究结果强调，针对性遗忘是保护生成模型的一种有前景的防御手段。代码和模型权重可在https://github.com/Mystic-Slice/Sealing-The-Backdoor获取。
英文摘要：
Text-to-image diffusion models have revolutionized generative AI, but their vulnerability to backdoor attacks poses significant security risks. Adversaries can inject imperceptible textual triggers into training data, causing models to generate manipulated outputs. Although text-based backdoor defenses in classification models are well-explored, generative models lack effective mitigation techniques against. We address this by selectively erasing the model's learned associations between adversarial text triggers and poisoned outputs, while preserving overall generation quality. Our approach, Self-Knowledge Distillation with Cross-Attention Guidance (SKD-CAG), uses knowledge distillation to guide the model in correcting responses to poisoned prompts while maintaining image quality by exploiting the fact that the backdoored model still produces clean outputs in the absence of triggers. Using the cross-attention mechanism, SKD-CAG neutralizes backdoor influences at the attention level, ensuring the targeted removal of adversarial effects. Extensive experiments show that our method outperforms existing approaches, achieving removal accuracy 100\% for pixel backdoors and 93\% for style-based attacks, without sacrificing robustness or image fidelity. Our findings highlight targeted unlearning as a promising defense to secure generative models. Code and model weights can be found at https://github.com/Mystic-Slice/Sealing-The-Backdoor .
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。
