Online Continual Learning via the Knowledge Invariant and Spread-out Properties

TL;DR

This paper introduces a novel online continual learning method called OCLKISP that leverages knowledge invariant and spread-out properties to transfer structural knowledge and reduce catastrophic forgetting.

Contribution

The paper proposes a new approach that constrains embedding evolution using KISP, enabling transfer of inter-instance structural knowledge and mitigating learning bias in continual learning.

Findings

Outperforms state-of-the-art methods on four benchmarks.

Effectively transfers structural knowledge between tasks.

Reduces catastrophic forgetting in continual learning.

Abstract

The goal of continual learning is to provide intelligent agents that are capable of learning continually a sequence of tasks using the knowledge obtained from previous tasks while performing well on prior tasks. However, a key challenge in this continual learning paradigm is catastrophic forgetting, namely adapting a model to new tasks often leads to severe performance degradation on prior tasks. Current memory-based approaches show their success in alleviating the catastrophic forgetting problem by replaying examples from past tasks when new tasks are learned. However, these methods are infeasible to transfer the structural knowledge from previous tasks i.e., similarities or dissimilarities between different instances. Furthermore, the learning bias between the current and prior tasks is also an urgent problem that should be solved. In this work, we propose a new method, named Online Continual Learning via the Knowledge Invariant and Spread-out Properties (OCLKISP), in which we constrain the evolution of the embedding features via Knowledge Invariant and Spread-out Properties (KISP). Thus, we can further transfer the inter-instance structural knowledge of previous tasks while alleviating the forgetting due to the learning bias. We empirically evaluate our proposed method on four popular benchmarks for continual learning: Split CIFAR 100, Split SVHN, Split CUB200 and Split Tiny-Image-Net. The experimental results show the efficacy of our proposed method compared to the state-of-the-art continual learning algorithms.