New metrics for analyzing continual learners

TL;DR

This paper introduces new metrics for continual learning that consider the increasing difficulty of tasks, providing better insights into model stability and plasticity over sequential tasks.

Contribution

The paper proposes novel metrics for continual learning that account for task difficulty, addressing limitations of existing metrics and enhancing evaluation of stability-plasticity trade-offs.

Findings

Proposed metrics reveal new insights into model performance over time.

Existing metrics overlook task difficulty, leading to incomplete evaluations.

Experiments show improved understanding of stability-plasticity balance.

Abstract

Deep neural networks have shown remarkable performance when trained on independent and identically distributed data from a fixed set of classes. However, in real-world scenarios, it can be desirable to train models on a continuous stream of data where multiple classification tasks are presented sequentially. This scenario, known as Continual Learning (CL) poses challenges to standard learning algorithms which struggle to maintain knowledge of old tasks while learning new ones. This stability-plasticity dilemma remains central to CL and multiple metrics have been proposed to adequately measure stability and plasticity separately. However, none considers the increasing difficulty of the classification task, which inherently results in performance loss for any model. In that sense, we analyze some limitations of current metrics and identify the presence of setup-induced forgetting. Therefore, we propose new metrics that account for the task's increasing difficulty. Through experiments on benchmark datasets, we demonstrate that our proposed metrics can provide new insights into the stability-plasticity trade-off achieved by models in the continual learning environment.