arXiv cs.AI· Yiran Xiahou, Siliang Qin, Kai Chen, Shengzhi Zhang, Renyang Liu·· 12 小时前AI 评分34
Concept Acupuncturist：基于稀疏自编码器的扩散模型高效精准概念擦除方法
Concept Acupuncturist: An Efficient and Precise Concept Erasure Method for Diffusion Models via Sparse Autoencoders
AI 导读
研究者提出 ConceptACP，一个结合可解释性分析与免推理微调的扩散模型机器遗忘框架，通过将语义概念映射到 U-Net 交叉注意力块并定位最相关块进行定向优化，避免全扩散轨迹优化。该方法利用稀疏自编码器（SAE）构造特征级修正信号以缓解概念纠缠，无需外部推理模块。在 SDXL-Turbo 上，其标准单步生成的平均遗忘成功率较最佳基线提升 36.34 个百分点，并保持图像质量与提示词保真度。
正文 · 原文
Abstract:Diffusion-based image generation models produce high-quality images but may generate harmful or policy-violating content, motivating machine unlearning (MU) to selectively remove target concepts while preserving general capabilities. However, existing MU approaches often struggle with the distributed and entangled nature of concept representations, leading to inefficient optimization and imprecise edits when applied through large-scale fine-tuning over full diffusion trajectories. To bridge this gap, we propose Concept Acupuncturist (ConceptACP), a novel MU framework that integrates interpretability-guided analysis with inference-free fine-tuning for efficient and precise concept removal. By establishing mappings between semantic concepts and cross-attention blocks in the U-Net denoiser and localizing the most concept-relevant block for targeted optimization, ConceptACP avoids full diffusion trajectory optimization, thereby improving fine-tuning efficiency. To mitigate concept entanglement, we propose a feature-level correction mechanism based on Sparse Autoencoders (SAEs), where SAE-derived sparse, semantically meaningful feature activations are manipulated to construct targeted correction signals, which are then projected back to the output space of the corresponding block and used as supervision for precise concept removal. Consequently, ConceptACP eliminates the need for external inference-time modules while achieving efficient and fine-grained unlearning. Instantiated on SDXL-Turbo, ConceptACP improves the forgetting success rate by 36.34 percentage points in the standard one-step generation setting on average over the best-performing baseline, while preserving image quality and prompt fidelity, and generalizes effectively across other Stable Diffusion backbones. Our code is available at this https URL.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.32059 [cs.CV] |
| (or arXiv:2610.32059v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.32059 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yiran Xiahou [view email]
[v1]
Fri, 23 Oct 2026 01:54:50 UTC (25,179 KB)
来源：arXiv cs.AI · arxiv.org