提出DeCoFlow，解决工业异常检测中持续学习时的遗忘问题。
DeCoFlow: Structural Decomposition of Normalizing Flows for Continual Anomaly Detection
- 将归一化流分解为固定基底与任务专用低秩适配器，隔离参数更新。
- 在MVTec-AD和VisA上分别达98.40%和93.00%图像级AUROC。
- 仅需每任务227万参数，实现零参数遗忘，适合资源受限场景。
工业环境中新产品类别按序出现，要求在无历史数据的情况下持续进行异常检测。归一化流（NFs）虽能提供精确密度估计，但在跨任务参数更新时易因灾难性遗忘导致密度流形扭曲。尽管参数隔离可避免干扰，但必须保持NF的严格可逆性与雅可比行列式有效性。我们利用仿射耦合层的内在特性——其变换有效性不依赖子网参数化形式，提出DeCoFlow：将子网分解为冻结的通用基底与任务特定的低秩适配器以实现更新隔离。进一步引入任务特定对齐、辅助耦合层与尾部感知损失，缓解固定基底刚性。DeCoFlow在MVTec-AD上达到98.40%图像级AUROC，VisA上达93.00%，且在正确路由下参数级遗忘为0.00%，每任务仅需227万参数。
原文摘要 · Abstract (English)
In industrial environments, new product categories arrive sequentially, requiring continual anomaly detection without access to past data. Normalizing Flows (NFs) provide exact density estimation but suffer from catastrophic forgetting as parameter updates across tasks distort the density manifold. While parameter isolation can prevent interference, it must preserve the strict invertibility and Jacobian validity of NFs. To satisfy these requirements, we exploit the inherent property that affine coupling layers maintain transformation validity regardless of subnet parameterization. Based on this, we propose DeCoFlow, which decomposes subnets into a frozen universal base and task-specific low-rank adapters to isolate updates. We further introduce Task-Specific Alignment, Auxiliary Coupling Layers, and Tail-Aware Loss to compensate for frozen-base rigidity. DeCoFlow achieves state-of-the-art image-level AUROCs of 98.40% on MVTec-AD and 93.00% on VisA, while maintaining parameter-level zero forgetting (0.00% FM under correct routing) with only 2.27M parameters per task.
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。