Optimal Task Order for Continual Learning of Multiple Tasks

Ziyan Li, Naoki Hiratani

Continual learning of multiple tasks remains a major challenge for neural networks. Here, we investigate how task order influences continual learning and propose a strategy for opt…