This work investigates the entanglement between Continual Learning (CL) and Transfer Learning (TL). In particular, we shed light on the widespread application of network pretraining, highlighting that it is itself subject to catastrophic forgetting. Unfortunately, this issue leads to the under-exploitation of knowledge transfer during later tasks. On this ground, we propose Transfer without Forgetting (TwF), a hybrid Continual Transfer Learning approach building upon a fixed pretrained sibling network, which continuously propagates the knowledge inherent in the source domain through a layer-wise loss term. Our experiments indicate that TwF steadily outperforms other CL methods across a variety of settings, averaging a 4.81% gain in Class-Incremental accuracy over a variety of datasets and different buffer sizes.

翻译：这项工作调查了持续学习和转移学习之间的纠缠关系。 特别是,我们揭示了网络预培训的广泛应用,强调网络预培训本身会受到灾难性的遗忘。 不幸的是,这个问题导致在以后的任务中知识转让的利用不足。 在此基础上,我们提议在不忘记的情况下转让(TwF),即基于固定的预先训练的辅助网络的混合持续转让学习方法,它通过一个分层损失术语不断传播源域内固有的知识。 我们的实验表明,TwF在各种环境中稳步地超越了其他CL方法,在各种数据集和不同的缓冲大小上平均增加了4.81%的等级偏差精确度。