面向弱监督视频异常检测的混合记忆专家联合学习通用与多样模式
Joint Learning of General and Diverse Patterns with Mixture of Memory Experts for Weakly-Supervised Video Anomaly Detection
提出混合记忆专家框架，每个专家有内部记忆做细粒度特化、共享外部记忆聚合通用知识，并用大语言模型构建异常原型引导专家路由，为视频异常检测中同时建模泛化性与多样性提供了可复用的路由与记忆机制。
原文依据 we propose the Mixture of Memory Experts (MoME), a unified framework that jointly learns general and diverse patterns
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论弱监督视频异常检测（wVAD）旨在仅使用二值标签检测异常事件，这使得同时捕捉异常的多样性及其共享语义线索具有挑战性。现有方法要么聚焦于通用异常模式，虽泛化性强但判别力弱；要么依赖类别级多样性建模，忽略了共享语义且泛化能力有限。为克服这些局限，我们提出混合记忆专家（MoME），一个联合学习通用与多样模式的统一框架。MoME中的每个专家拥有用于细粒度特化的内部记忆，并共享用于通用知识聚合的外部记忆。为增强语义多样性并提升超越粗粒度类别级监督的泛化能力，我们引入异常原型路由器，利用大语言模型构建泛化的异常原型以进行语义引导的专家路由。此外，通过针对APR的正则化损失、针对专家的差异性损失以及重建与记忆任务，分别确保了均衡路由、专家多样性和增强的模式判别力。在UCF-Crime和XD-Violence上的大量实验表明，我们的方法达到了最优性能，验证了联合建模通用性与多样性对于鲁棒异常检测的有效性。
英文原文摘要 1443 字符
Weakly-supervised Video Anomaly Detection (wVAD) aims to detect abnormal events using only binary labels, making it challenging to capture both the diversity of anomalies and their shared semantic cues. Existing methods either focus on a generic anomaly pattern, achieving strong generalization but weak discrimination, or rely on class-level diversity modeling, which ignores shared semantics and suffers from limited generalization. To overcome these limitations, we propose the Mixture of Memory Experts (MoME), a unified framework that jointly learns general and diverse patterns. Each expert in MoME possesses an internal memory for fine-grained specialization and shares an external memory for general knowledge aggregation. To enhance semantic diversity and improve generalization beyond coarse class-level supervision, we introduce an Anomaly Prototype Router that leverages large language models to construct generalized anomaly prototypes for semantically guided expert routing. Moreover, balanced routing, expert diversity, and enhanced pattern discriminability are ensured by the regularization loss for APR, the distinctiveness loss for experts, and reconstruction together with memory tasks, respectively. Extensive experiments on UCF-Crime and XD-Violence demonstrate that our approach achieves state-of-the-art performance, validating the effectiveness of jointly modeling generality and diversity for robust anomaly detection.