Exemplar-free Online Continual Learning

TL;DR

This paper introduces an exemplar-free online continual learning method using a nearest-class-mean classifier, which outperforms exemplar-based methods on image classification benchmarks without storing data exemplars.

Contribution

The work proposes a novel exemplar-free approach leveraging online class mean estimation, reducing storage needs and privacy concerns in continual learning.

Findings

Outperforms state-of-the-art exemplar-based methods on CIFAR-100 and Food-1k.

Achieves competitive performance with larger exemplar sizes.

Effective in online learning scenarios with single-pass data.

Abstract

Targeted for real world scenarios, online continual learning aims to learn new tasks from sequentially available data under the condition that each data is observed only once by the learner. Though recent works have made remarkable achievements by storing part of learned task data as exemplars for knowledge replay, the performance is greatly relied on the size of stored exemplars while the storage consumption is a significant constraint in continual learning. In addition, storing exemplars may not always be feasible for certain applications due to privacy concerns. In this work, we propose a novel exemplar-free method by leveraging nearest-class-mean (NCM) classifier where the class mean is estimated during training phase on all data seen so far through online mean update criteria. We focus on image classification task and conduct extensive experiments on benchmark datasets including CIFAR-100 and Food-1k. The results demonstrate that our method without using any exemplar outperforms state-of-the-art exemplar-based approaches with large margins under standard protocol (20 exemplars per class) and is able to achieve competitive performance even with larger exemplar size (100 exemplars per class).