Fine-Grained Knowledge Selection and Restoration for Non-Exemplar Class Incremental Learning
This addresses the problem of learning new tasks without access to old data for machine learning systems, representing an incremental improvement over existing knowledge distillation methods.
The paper tackles catastrophic forgetting in non-exemplar class incremental learning by proposing a framework with fine-grained selective patch-level distillation and task-agnostic prototype generation, achieving state-of-the-art results on datasets like CIFAR100 and ImageNet-Subset.
Non-exemplar class incremental learning aims to learn both the new and old tasks without accessing any training data from the past. This strict restriction enlarges the difficulty of alleviating catastrophic forgetting since all techniques can only be applied to current task data. Considering this challenge, we propose a novel framework of fine-grained knowledge selection and restoration. The conventional knowledge distillation-based methods place too strict constraints on the network parameters and features to prevent forgetting, which limits the training of new tasks. To loose this constraint, we proposed a novel fine-grained selective patch-level distillation to adaptively balance plasticity and stability. Some task-agnostic patches can be used to preserve the decision boundary of the old task. While some patches containing the important foreground are favorable for learning the new task. Moreover, we employ a task-agnostic mechanism to generate more realistic prototypes of old tasks with the current task sample for reducing classifier bias for fine-grained knowledge restoration. Extensive experiments on CIFAR100, TinyImageNet and ImageNet-Subset demonstrate the effectiveness of our method. Code is available at https://github.com/scok30/vit-cil.