Better Knowledge Retention through Metric Learning

Ke Li*, Shichong Peng*, Kailas Vodrahalli*, Jitendra Malik

TL;DR: We show metric learning can help reduce catastrophic forgetting

Abstract: In a continual learning setting, new categories may be introduced over time, and an ideal learning system should perform well on both the original categories and the new categories. While deep neural nets have achieved resounding success in the classical setting, they are known to forget about knowledge acquired in prior episodes of learning if the examples encountered in the current episode of learning are drastically different from those encountered in prior episodes. This makes deep neural nets ill-suited to continual learning. In this paper, we propose a new model that can both leverage the expressive power of deep neural nets and is resilient to forgetting when new categories are introduced. We demonstrate an improvement in terms of accuracy on original classes compared to a vanilla deep neural net.

Keywords: metric learning, continual learning, catastrophic forgetting