Generalizable Two-Branch Framework for Image Class-Incremental Learning

arXiv:2402.18086v1 Announce Type: new
Abstract: Deep neural networks often severely forget previously learned knowledge when learning new knowledge. Various continual learning (CL) methods have been proposed to handle such a catastrophic forgetting issue from different perspectives and achieved substantial improvements.In this paper, a novel two-branch continual learning framework is proposed to further enhance most existing CL methods. Specifically, the main branch can be any existing CL model and the newly introduced side branch is a lightweight convolutional network. The output …

abstract arxiv catastrophic forgetting class continual cs.cv framework image improvements incremental issue knowledge networks neural networks novel paper perspectives type