Diffusion-Driven Data Replay: A Novel Approach to Combat Forgetting in Federated Class Continual Learning

TL;DR

This paper introduces a diffusion model-based data replay technique to mitigate catastrophic forgetting in federated class continual learning, leveraging pre-trained diffusion models for efficient class-specific data generation and improved model performance.

Contribution

It proposes a novel diffusion model-based data replay method for FCCL, reducing computational costs and enhancing class representation without training generative models from scratch.

Findings

Outperforms existing FCCL methods in experiments

Reduces training time and computational resources

Enhances classifier generalization through contrastive learning

Abstract

Federated Class Continual Learning (FCCL) merges the challenges of distributed client learning with the need for seamless adaptation to new classes without forgetting old ones. The key challenge in FCCL is catastrophic forgetting, an issue that has been explored to some extent in Continual Learning (CL). However, due to privacy preservation requirements, some conventional methods, such as experience replay, are not directly applicable to FCCL. Existing FCCL methods mitigate forgetting by generating historical data through federated training of GANs or data-free knowledge distillation. However, these approaches often suffer from unstable training of generators or low-quality generated data, limiting their guidance for the model. To address this challenge, we propose a novel method of data replay based on diffusion models. Instead of training a diffusion model, we employ a pre-trained conditional diffusion model to reverse-engineer each class, searching the corresponding input conditions for each class within the model's input space, significantly reducing computational resources and time consumption while ensuring effective generation. Furthermore, we enhance the classifier's domain generalization ability on generated and real data through contrastive learning, indirectly improving the representational capability of generated data for real data. Comprehensive experiments demonstrate that our method significantly outperforms existing baselines. Code is available at https://github.com/jinglin-liang/DDDR.