Continual Learning via Manifold Expansion Replay
This addresses the problem of catastrophic forgetting for continual learning systems, offering an incremental improvement over existing rehearsal-based methods.
The paper tackles catastrophic forgetting in continual learning by proposing Manifold Expansion Replay (MaER), a novel replay strategy that expands the implicit manifold of knowledge representations and uses Wasserstein distance for distillation, resulting in significant accuracy improvements on datasets like MNIST, CIFAR10, CIFAR100, and TinyImageNet, outperforming state-of-the-art methods.
In continual learning, the learner learns multiple tasks in sequence, with data being acquired only once for each task. Catastrophic forgetting is a major challenge to continual learning. To reduce forgetting, some existing rehearsal-based methods use episodic memory to replay samples of previous tasks. However, in the process of knowledge integration when learning a new task, this strategy also suffers from catastrophic forgetting due to an imbalance between old and new knowledge. To address this problem, we propose a novel replay strategy called Manifold Expansion Replay (MaER). We argue that expanding the implicit manifold of the knowledge representation in the episodic memory helps to improve the robustness and expressiveness of the model. To this end, we propose a greedy strategy to keep increasing the diameter of the implicit manifold represented by the knowledge in the buffer during memory management. In addition, we introduce Wasserstein distance instead of cross entropy as distillation loss to preserve previous knowledge. With extensive experimental validation on MNIST, CIFAR10, CIFAR100, and TinyImageNet, we show that the proposed method significantly improves the accuracy in continual learning setup, outperforming the state of the arts.