Recent, rapid advancement in visual question answering architecture: a review

TL;DR

This review paper summarizes recent rapid advancements in visual question answering architectures, emphasizing the growth of multimodal systems and their significance in AI over the past few years.

Contribution

It provides an updated overview of recent developments in visual question answering architectures, building on previous reviews and highlighting new trends and improvements.

Findings

Significant growth in research on multimodal architectures

Recent improvements in VQA system accuracy and efficiency

Increased importance of VQA in AI applications

Abstract

Understanding visual question answering is going to be crucial for numerous human activities. However, it presents major challenges at the heart of the artificial intelligence endeavor. This paper presents an update on the rapid advancements in visual question answering using images that have occurred in the last couple of years. Tremendous growth in research on improving visual question answering system architecture has been published recently, showing the importance of multimodal architectures. Several points on the benefits of visual question answering are mentioned in the review paper by Manmadhan et al. (2020), on which the present article builds, including subsequent updates in the field.