Generative Negative Replay for Continual Learning

TL;DR

This paper introduces a novel generative negative replay method for continual learning, which uses generated data as negative examples to improve learning of new classes, especially in complex, high-dimensional scenarios.

Contribution

It proposes a new approach where generative models provide negative replay data to enhance learning of new classes in continual learning, addressing limitations of traditional generative replay.

Findings

Effective in complex class-incremental scenarios

Improves learning of new classes with small data experiences

Outperforms existing generative replay methods in high-dimensional data

Abstract

Learning continually is a key aspect of intelligence and a necessary ability to solve many real-life problems. One of the most effective strategies to control catastrophic forgetting, the Achilles' heel of continual learning, is storing part of the old data and replaying them interleaved with new experiences (also known as the replay approach). Generative replay, which is using generative models to provide replay patterns on demand, is particularly intriguing, however, it was shown to be effective mainly under simplified assumptions, such as simple scenarios and low-dimensional data. In this paper, we show that, while the generated data are usually not able to improve the classification accuracy for the old classes, they can be effective as negative examples (or antagonists) to better learn the new classes, especially when the learning experiences are small and contain examples of just one or few classes. The proposed approach is validated on complex class-incremental and data-incremental continual learning scenarios (CORe50 and ImageNet-1000) composed of high-dimensional data and a large number of training experiences: a setup where existing generative replay approaches usually fail.