Prototypes-Injected Prompt for Federated Class Incremental Learning

Muhammad Anwar Ma'sum, Mahardhika Pratama, Savitha Ramasamy, Lin Liu, H Habibullah, Ryszard Kowalczyk

Keywords: federated class incremental learning, federated learning, class incremental learning, continual learning, prompt, prototype

Abstract: Federated Class Incremental Learning (FCIL) is a new challenge in continual learning (CL) that addresses catastrophic forgetting and non-IID data distribution simultaneously. Existing FCIL methods call for high communication costs and exemplars from previous classes along with performance issues. 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 results show that the proposed method outperforms the current state of the arts (SOTAs) with a significant gap of 14-33% in CIFAR100, MiniImageNet, and TinyImageNet datasets. Our extensive analysis demonstrates the robustness of our proposed method in different task sizes, small participating local clients, and small global rounds. For further study, source codes of PIP, baseline, and experimental logs are shared publicly in https://anonymous.4open.science/r/an122pouyyt789/.

Primary Area: transfer learning, meta learning, and lifelong learning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 9261