通过正交投影实现持续机器遗忘，避免参数干扰。
Orthogonal Subspace Projection for Continual Machine Unlearning via SVD-Based LoRA
- 用SVD引导的正交子空间投影约束每次更新方向。
- 30次连续删除后保留准确率仍达58.1%，删除效果强。
- 适合需要长期安全数据清除的模型部署场景。
持续机器遗忘旨在移除不应再保留的数据影响，同时保持模型在其他任务上的有效性。当删除请求按顺序到来时，模型需反复适应而不丢失已有知识，这极具挑战性。低秩适配（LoRA）可高效实现此类更新，但直接组合多个序列化LoRA模块会导致参数冲突，引发任务间强烈干扰。本文提出一种基于奇异值分解（SVD）的静态正交子空间投影方法：在训练中约束每个新LoRA更新位于先前任务子空间的正交补空间中，从而实现任务隔离且无需部署时动态路由。在CIFAR-100（ResNet-20）和MNIST上的实验表明，该方法在长序列遗忘任务中表现稳定。经过30次连续遗忘任务后，现有最优静态融合方法将保留准确率从60.39%降至12.70%，而本文方法在维持约58.1%基线性能的同时，仍具备强遗忘能力。
原文摘要 · Abstract (English)
Continual machine unlearning aims to remove the influence of data that should no longer be retained, while preserving the usefulness of the model on everything else. This setting becomes especially difficult when deletion requests arrive sequentially, because the model must repeatedly adapt without erasing previously retained knowledge. Low-Rank Adaptation (LoRA) offers an efficient way to implement such updates, but naively combining many sequential LoRA modules leads to parameter collision, causing \textit{strong interference} between tasks. We propose a static alternative based on Singular Value Decomposition (SVD)-guided orthogonal subspace projection. Our method constrains each new LoRA update during training so that it lies in the orthogonal complement of the subspaces used by earlier unlearning tasks. This preserves task isolation without requiring dynamic routing at deployment. Experiments on CIFAR-100 with ResNet-20 and on MNIST show stable behavior across long sequences of unlearning tasks. After thirty sequential unlearning tasks, state-of-the-art static fusion reduces retained accuracy from 60.39\% to 12.70\%, whereas the proposed in-training constrained optimization maintains baseline performance ($\sim$58.1\%) while preserving strong unlearning efficacy.
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。