Deep Active Learning for Computer Vision: Past and Future

TL;DR

This paper reviews deep active learning in computer vision, highlighting recent advancements, applications, industrial uses, limitations, and future directions to enhance AI model development and democratization.

Contribution

It provides a comprehensive overview of deep active learning techniques, applications, and challenges, emphasizing its importance in scalable AI model production.

Findings

Active learning improves data efficiency in deep neural networks.

Applications in computer vision demonstrate practical benefits.

Identifies limitations and future research directions in the field.

Abstract

As an important data selection schema, active learning emerges as the essential component when iterating an Artificial Intelligence (AI) model. It becomes even more critical given the dominance of deep neural network based models, which are composed of a large number of parameters and data hungry, in application. Despite its indispensable role for developing AI models, research on active learning is not as intensive as other research directions. In this paper, we present a review of active learning through deep active learning approaches from the following perspectives: 1) technical advancements in active learning, 2) applications of active learning in computer vision, 3) industrial systems leveraging or with potential to leverage active learning for data iteration, 4) current limitations and future research directions. We expect this paper to clarify the significance of active learning in a modern AI model manufacturing process and to bring additional research attention to active learning. By addressing data automation challenges and coping with automated machine learning systems, active learning will facilitate democratization of AI technologies by boosting model production at scale.