work MoRAM Little By Little: Continual Learning via Incremental Mixture of Rank-1 Associative Memory Experts
