Recent Advances of Continual Learning in Computer Vision: An Overview

TL;DR

This paper provides a comprehensive overview of recent progress in continual learning within computer vision, categorizing techniques like regularization, knowledge distillation, memory, and generative replay, and discussing future research directions.

Contribution

It offers a structured review of recent methods in continual learning for computer vision, highlighting their characteristics, applications, and potential future subareas.

Findings

Categorization of techniques: regularization, distillation, memory, generative replay.

Analysis of each technique's characteristics and applications.

Discussion of underexplored subareas in continual learning.

Abstract

In contrast to batch learning where all training data is available at once, continual learning represents a family of methods that accumulate knowledge and learn continuously with data available in sequential order. Similar to the human learning process with the ability of learning, fusing, and accumulating new knowledge coming at different time steps, continual learning is considered to have high practical significance. Hence, continual learning has been studied in various artificial intelligence tasks. In this paper, we present a comprehensive review of the recent progress of continual learning in computer vision. In particular, the works are grouped by their representative techniques, including regularization, knowledge distillation, memory, generative replay, parameter isolation, and a combination of the above techniques. For each category of these techniques, both its characteristics and applications in computer vision are presented. At the end of this overview, several subareas, where continuous knowledge accumulation is potentially helpful while continual learning has not been well studied, are discussed.