A Review of Recent Advances of Binary Neural Networks for Edge Computing

TL;DR

This paper reviews recent advances in binary neural networks and 1-bit CNNs, emphasizing their suitability for edge computing in applications like vision and speech recognition, and discusses future prospects.

Contribution

It provides a comprehensive classification and summary of recent BNN and 1-bit CNN research, highlighting their architectures, training methods, and applications for edge computing.

Findings

BNNs are effective for real-time edge applications.

Recent methods improve accuracy and efficiency of BNNs.

BNNs are increasingly applied in vision and speech tasks.

Abstract

Edge computing is promising to become one of the next hottest topics in artificial intelligence because it benefits various evolving domains such as real-time unmanned aerial systems, industrial applications, and the demand for privacy protection. This paper reviews recent advances on binary neural network (BNN) and 1-bit CNN technologies that are well suitable for front-end, edge-based computing. We introduce and summarize existing work and classify them based on gradient approximation, quantization, architecture, loss functions, optimization method, and binary neural architecture search. We also introduce applications in the areas of computer vision and speech recognition and discuss future applications for edge computing.