continual learning is a challenging problem in which models need to be
trained on non-stationary data across sequential tasks for class-incremental
learning. While previous methods have focused on using either regularization or
rehearsal-based frameworks to alleviate catastrophic forge
本文提出一种基于 EM 框架的领域感知的持续学习方法来应对在增量学习中类分布和域分布的变化问题,使用基于 von Mises-Fisher 混合模型的灵活类表示来捕获类内结构,采用内部平衡和交叉类数据的处理方式来设计双层平衡记忆,并结合蒸馏损失实现更好的稳定性和可塑性权衡。经过大量实验,表明该方法在三个基准数据集上均优于现有方法。