Compacting Deep Neural Networks for Internet of Things: Methods and Applications
It addresses the problem of high computational and storage costs hindering DNN deployment on IoT devices, but it is incremental as it is a survey paper summarizing existing technologies.
This paper presents a comprehensive survey of technologies for compacting deep neural networks (DNNs) to enable their deployment on resource-constrained Internet of Things (IoT) devices, categorizing methods into network model compression, knowledge distillation, and modification of network structures, and discussing applications and future directions.
Deep Neural Networks (DNNs) have shown great success in completing complex tasks. However, DNNs inevitably bring high computational cost and storage consumption due to the complexity of hierarchical structures, thereby hindering their wide deployment in Internet-of-Things (IoT) devices, which have limited computational capability and storage capacity. Therefore, it is a necessity to investigate the technologies to compact DNNs. Despite tremendous advances in compacting DNNs, few surveys summarize compacting-DNNs technologies, especially for IoT applications. Hence, this paper presents a comprehensive study on compacting-DNNs technologies. We categorize compacting-DNNs technologies into three major types: 1) network model compression, 2) Knowledge Distillation (KD), 3) modification of network structures. We also elaborate on the diversity of these approaches and make side-by-side comparisons. Moreover, we discuss the applications of compacted DNNs in various IoT applications and outline future directions.