Most artificial intelligence models have limiting ability to solve new tasks
faster, without forgetting previously acquired knowledge. The recently emerging
paradigm of continual learning aims to solve this issue, in which the model
learns various tasks in a sequential fashion. In this