Beyond Supervised Continual Learning: a Review

TL;DR

This review explores continual learning beyond supervised classification, including unsupervised, semi-supervised, and reinforcement learning, highlighting unique challenges and potential contributions in these diverse settings.

Contribution

It introduces a schema for classifying CL approaches based on autonomy and supervision levels, and discusses challenges and opportunities in non-supervised CL.

Findings

Most CL research focuses on supervised learning.

Different settings like unsupervised and reinforcement learning face unique challenges.

The paper proposes a classification schema for CL approaches.

Abstract

Continual Learning (CL, sometimes also termed incremental learning) is a flavor of machine learning where the usual assumption of stationary data distribution is relaxed or omitted. When naively applying, e.g., DNNs in CL problems, changes in the data distribution can cause the so-called catastrophic forgetting (CF) effect: an abrupt loss of previous knowledge. Although many significant contributions to enabling CL have been made in recent years, most works address supervised (classification) problems. This article reviews literature that study CL in other settings, such as learning with reduced supervision, fully unsupervised learning, and reinforcement learning. Besides proposing a simple schema for classifying CL approaches w.r.t. their level of autonomy and supervision, we discuss the specific challenges associated with each setting and the potential contributions to the field of CL in general.