如有侵权或其他问题欢迎留言联系更正或删除。出处IJCAI 2024尚未开源一. 绪论1.Traditional statistical methoddo not give good results onhigh dimensionality, being unstructured and complexmultivariate time series due to theirheavy reliance on a prior knowledge (严重的先验依赖).2. DL-TSAD 核心思想对正常模式、行为进行隐式建模 (implicitly modeling normal patterns and behaviors)3. Fortime series data augmentation, the common practice is tointroduce perturbationsto the original sequences. Given the unique sequential and periodic traits of time series, various augmentation techniques are explored. However, existing methods often prioritize distinguishing between normal and abnormal instances,which leads to a coarse classification that may result in the problem that the model is overly sensitive to the choice of perturbation, requires a lot of experimentation to try out the appropriate perturbation.4. 对比学习 挑战正、负样本的选择可能导致来自同一类的样本被错误地视为负样本从而导致类崩溃和无效的特征表示5.核心思想the higher the probability of misclassifying perturbation types, the higher the probability of anomalies.二. Method投影网络分支举一个具体的例子说明三. 实验结果
