Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey on Collaborative DNN Inference for Edge Intelligence

Weiqing Ren, Yuben Qu, Chao Dong, Yuqian Jing, Hao Sun, Qihui Wu, Song Guo

TL;DR

Cloud-only DNN inference on edge data incurs high latency and privacy risks. This paper surveys EI-oriented collaborative DNN inference across four paradigms—cloud-device, edge-device, cloud-edge-device, and device-device—providing taxonomy, architectures, optimization techniques, and performance insights. It synthesizes challenges and proposes future directions in dynamic networks, intelligent task allocation, and security/privacy to guide practical EI deployments. The work serves as a structured reference for researchers and practitioners designing scalable, efficient, and privacy-preserving EI systems.

Abstract

With the vigorous development of artificial intelligence (AI), the intelligent applications based on deep neural network (DNN) change people's lifestyles and the production efficiency. However, the huge amount of computation and data generated from the network edge becomes the major bottleneck, and traditional cloud-based computing mode has been unable to meet the requirements of real-time processing tasks. To solve the above problems, by embedding AI model training and inference capabilities into the network edge, edge intelligence (EI) becomes a cutting-edge direction in the field of AI. Furthermore, collaborative DNN inference among the cloud, edge, and end device provides a promising way to boost the EI. Nevertheless, at present, EI oriented collaborative DNN inference is still in its early stage, lacking a systematic classification and discussion of existing research efforts. Thus motivated, we have made a comprehensive investigation on the recent studies about EI oriented collaborative DNN inference. In this paper, we firstly review the background and motivation of EI. Then, we classify four typical collaborative DNN inference paradigms for EI, and analyze the characteristics and key technologies of them. Finally, we summarize the current challenges of collaborative DNN inference, discuss the future development trend and provide the future research direction.

A Survey on Collaborative DNN Inference for Edge Intelligence

TL;DR

Cloud-only DNN inference on edge data incurs high latency and privacy risks. This paper surveys EI-oriented collaborative DNN inference across four paradigms—cloud-device, edge-device, cloud-edge-device, and device-device—providing taxonomy, architectures, optimization techniques, and performance insights. It synthesizes challenges and proposes future directions in dynamic networks, intelligent task allocation, and security/privacy to guide practical EI deployments. The work serves as a structured reference for researchers and practitioners designing scalable, efficient, and privacy-preserving EI systems.

Abstract

With the vigorous development of artificial intelligence (AI), the intelligent applications based on deep neural network (DNN) change people's lifestyles and the production efficiency. However, the huge amount of computation and data generated from the network edge becomes the major bottleneck, and traditional cloud-based computing mode has been unable to meet the requirements of real-time processing tasks. To solve the above problems, by embedding AI model training and inference capabilities into the network edge, edge intelligence (EI) becomes a cutting-edge direction in the field of AI. Furthermore, collaborative DNN inference among the cloud, edge, and end device provides a promising way to boost the EI. Nevertheless, at present, EI oriented collaborative DNN inference is still in its early stage, lacking a systematic classification and discussion of existing research efforts. Thus motivated, we have made a comprehensive investigation on the recent studies about EI oriented collaborative DNN inference. In this paper, we firstly review the background and motivation of EI. Then, we classify four typical collaborative DNN inference paradigms for EI, and analyze the characteristics and key technologies of them. Finally, we summarize the current challenges of collaborative DNN inference, discuss the future development trend and provide the future research direction.
Paper Structure (37 sections, 8 figures, 6 tables)

This paper contains 37 sections, 8 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Preliminaries and classification of collaborative DNN inference
  3. Introduction to EI
  4. Collaborative DNN inference for EI
  5. Various collaborative DNN inference paradigms
  6. Cloud-device collaborative DNN inference
  7. Motivation and architecture
  8. Performance optimization
  9. Total inference latency minimization
  10. Energy consumption minimization
  11. Summary and analysis
  12. Edge-device collaborative DNN inference
  13. Motivation and architecture
  14. Performance optimization
  15. Total inference latency minimization
  16. ...and 22 more sections

Figures (8)

  • Figure 1: An illustration of the existing cloud-edge-device network structure zhou2019edge.
  • Figure 2: Classification of EI.
  • Figure 3: Illustration of cloud-device collaborative DNN inference.
  • Figure 4: Illustration of edge-device collaborative DNN inference.
  • Figure 5: Illustration of early-exit mechanism in branchy AlexNet model. teerapittayanon2016branchynet
  • ...and 3 more figures