A Benchmark and Empirical Analysis for Replay Strategies in Continual Learning

TL;DR

This paper provides a comprehensive benchmark and empirical analysis of replay strategies in continual learning, evaluating their efficiency, performance, and scalability across multiple datasets and domains.

Contribution

It offers an in-depth evaluation of memory replay methods, comparing sampling strategies and proposing practical guidelines for selecting replay methods based on data distributions.

Findings

Replay strategies vary in efficiency and scalability.

Sampling strategies significantly impact performance.

Practical guidelines improve replay method selection.

Abstract

With the capacity of continual learning, humans can continuously acquire knowledge throughout their lifespan. However, computational systems are not, in general, capable of learning tasks sequentially. This long-standing challenge for deep neural networks (DNNs) is called catastrophic forgetting. Multiple solutions have been proposed to overcome this limitation. This paper makes an in-depth evaluation of the memory replay methods, exploring the efficiency, performance, and scalability of various sampling strategies when selecting replay data. All experiments are conducted on multiple datasets under various domains. Finally, a practical solution for selecting replay methods for various data distributions is provided.