Abstract: Deep neural networks (DNNs) struggle to learn in dy-namic settings because they mainly rely on static datasets. Continual learning (CL) aims to overcome this limitation by enabling DNNs to ...