Skip to main navigation Skip to search Skip to main content

Poster Abstract: E4: Energy-Efficient Early-Exit DNN Inference Framework for Edge Video Analytics

  • Harbin Institute of Technology
  • Harbin Institute of Technology Shenzhen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Deep neural networks (DNNs) are becoming extremely popular in video analytics applications at the edge. However, compute-intensive DNNs pose new challenges to achieve energy-efficient DNN inference on resource-constrained edge devices. In this paper, we propose E4, an energy-efficient DNN inference framework for edge video analytics. First, E4 analyzes video frame complexity by employing an attention-based cascade module that automatically determines DNN exit points. Second, E4's just-in-time (JIT) profiler leverages coordinate descent search to co-optimize the CPU and GPU clock frequencies for each layer before the DNN exit point. Preliminary experimental results show that E4 outperforms exiting methods in terms of power consumption and inference latency.

Original languageEnglish
Title of host publicationSenSys 2023 - Proceedings of the 21st ACM Conference on Embedded Networked Sensors Systems
PublisherAssociation for Computing Machinery, Inc
Pages512-513
Number of pages2
ISBN (Electronic)9798400704147
DOIs
StatePublished - 26 Apr 2024
Externally publishedYes
Event21st ACM Conference on Embedded Networked Sensors Systems, SenSys 2023 - Istanbul, Turkey
Duration: 13 Nov 202315 Nov 2023

Publication series

NameSenSys 2023 - Proceedings of the 21st ACM Conference on Embedded Networked Sensors Systems

Conference

Conference21st ACM Conference on Embedded Networked Sensors Systems, SenSys 2023
Country/TerritoryTurkey
CityIstanbul
Period13/11/2315/11/23

Keywords

  • DVFS
  • early-exit DNN
  • edge video analytics

Fingerprint

Dive into the research topics of 'Poster Abstract: E4: Energy-Efficient Early-Exit DNN Inference Framework for Edge Video Analytics'. Together they form a unique fingerprint.

Cite this