PIP: Prototypes-Injected Prompt for Federated Class Incremental Learning
This addresses the challenge of efficient and robust continual learning in federated settings for applications like distributed AI, though it appears incremental as it builds on existing prompt learning and prototype methods.
The paper tackles the problem of catastrophic forgetting and non-IID data distribution in Federated Class Incremental Learning (FCIL) by proposing a rehearsal-free method called prototypes-injected prompt (PIP), which achieves up to 33% improvement over state-of-the-art methods on datasets like CIFAR100, MiniImageNet, and TinyImageNet.
Federated Class Incremental Learning (FCIL) is a new direction in continual learning (CL) for addressing catastrophic forgetting and non-IID data distribution simultaneously. Existing FCIL methods call for high communication costs and exemplars from previous classes. We propose a novel rehearsal-free method for FCIL named prototypes-injected prompt (PIP) that involves 3 main ideas: a) prototype injection on prompt learning, b) prototype augmentation, and c) weighted Gaussian aggregation on the server side. Our experiment result shows that the proposed method outperforms the current state of the arts (SOTAs) with a significant improvement (up to 33%) in CIFAR100, MiniImageNet and TinyImageNet datasets. Our extensive analysis demonstrates the robustness of PIP in different task sizes, and the advantage of requiring smaller participating local clients, and smaller global rounds. For further study, source codes of PIP, baseline, and experimental logs are shared publicly in https://github.com/anwarmaxsum/PIP.