Data-Free Federated Class Incremental Learning with Diffusion-Based Generative Memory

TL;DR

This paper proposes a novel data-free federated class incremental learning framework using diffusion models to generate high-quality images, effectively mitigating catastrophic forgetting without extra communication costs.

Contribution

It introduces a diffusion-based generative memory, a balanced sampler, and an entropy-based sample filtering technique for improved federated class incremental learning.

Findings

Achieves over 4% higher average accuracy on Tiny-ImageNet.

Outperforms existing baselines across multiple datasets.

Effectively mitigates catastrophic forgetting in federated learning.

Abstract

Federated Class Incremental Learning (FCIL) is a critical yet largely underexplored issue that deals with the dynamic incorporation of new classes within federated learning (FL). Existing methods often employ generative adversarial networks (GANs) to produce synthetic images to address privacy concerns in FL. However, GANs exhibit inherent instability and high sensitivity, compromising the effectiveness of these methods. In this paper, we introduce a novel data-free federated class incremental learning framework with diffusion-based generative memory (DFedDGM) to mitigate catastrophic forgetting by generating stable, high-quality images through diffusion models. We design a new balanced sampler to help train the diffusion models to alleviate the common non-IID problem in FL, and introduce an entropy-based sample filtering technique from an information theory perspective to enhance the quality of generative samples. Finally, we integrate knowledge distillation with a feature-based regularization term for better knowledge transfer. Our framework does not incur additional communication costs compared to the baseline FedAvg method. Extensive experiments across multiple datasets demonstrate that our method significantly outperforms existing baselines, e.g., over a 4% improvement in average accuracy on the Tiny-ImageNet dataset.