Improving Replay Sample Selection and Storage for Less Forgetting in Continual Learning

TL;DR

This paper improves replay sample selection and storage strategies in continual learning to reduce forgetting, by comparing sampling methods and analyzing optimal sample quantities.

Contribution

It introduces a comparison of reservoir sampling with alternative strategies and analyzes how to determine the optimal number of stored samples.

Findings

Reservoir sampling is compared with other population strategies.

Analysis of the optimal number of samples to store.

Insights into reducing catastrophic forgetting.

Abstract

Continual learning seeks to enable deep learners to train on a series of tasks of unknown length without suffering from the catastrophic forgetting of previous tasks. One effective solution is replay, which involves storing few previous experiences in memory and replaying them when learning the current task. However, there is still room for improvement when it comes to selecting the most informative samples for storage and determining the optimal number of samples to be stored. This study aims to address these issues with a novel comparison of the commonly used reservoir sampling to various alternative population strategies and providing a novel detailed analysis of how to find the optimal number of stored samples.