Class-Incremental Learning using Diffusion Model for Distillation and Replay

TL;DR

This paper introduces a novel class-incremental learning approach that leverages a pretrained diffusion model to generate synthetic data for distillation and replay, effectively reducing catastrophic forgetting.

Contribution

It proposes using a diffusion model for synthetic data generation to enhance class-incremental learning, outperforming methods relying on external real datasets.

Findings

Improved accuracy on CIFAR100, ImageNet-Subset, and ImageNet benchmarks.

Synthetic data effectively aids in distillation and replay processes.

Outperforms existing state-of-the-art methods in incremental learning.

Abstract

Class-incremental learning aims to learn new classes in an incremental fashion without forgetting the previously learned ones. Several research works have shown how additional data can be used by incremental models to help mitigate catastrophic forgetting. In this work, following the recent breakthrough in text-to-image generative models and their wide distribution, we propose the use of a pretrained Stable Diffusion model as a source of additional data for class-incremental learning. Compared to competitive methods that rely on external, often unlabeled, datasets of real images, our approach can generate synthetic samples belonging to the same classes as the previously encountered images. This allows us to use those additional data samples not only in the distillation loss but also for replay in the classification loss. Experiments on the competitive benchmarks CIFAR100, ImageNet-Subset, and ImageNet demonstrate how this new approach can be used to further improve the performance of state-of-the-art methods for class-incremental learning on large scale datasets.