Layerwise Proximal Replay: A Proximal Point Method for Online Continual Learning

arXiv:2402.09542v1 Announce Type: new
Abstract: In online continual learning, a neural network incrementally learns from a non-i.i.d. data stream. Nearly all online continual learning methods employ experience replay to simultaneously prevent catastrophic forgetting and underfitting on past data. Our work demonstrates a limitation of this approach: networks trained with experience replay tend to have unstable optimization trajectories, impeding their overall accuracy. Surprisingly, these instabilities persist even when the replay buffer stores all previous training examples, suggesting that this issue is …

abstract arxiv catastrophic forgetting continual cs.lg data data stream experience network networks neural network type underfitting work