Dynamic Split Computing for Efficient Deep Edge Intelligence

Arian Bakhtiarnia, Nemanja Milošević, Qi Zhang, Dragana Bajović, Alexandros Iosifidis

Research output: Contribution to book/anthology/report/proceedingArticle in proceedingsResearchpeer-review


Deploying deep neural networks (DNNs) on IoT and mobile devices is a challenging task due to their limited computational resources. Thus, demanding tasks are often entirely offloaded to edge servers which can accelerate inference, however, it also causes communication cost and evokes privacy concerns. In addition, this approach leaves the computational capacity of end devices unused. Split computing is a paradigm where a DNN is split into two sections; the first section is executed on the end device, and the output is transmitted to the edge server where the final section is executed. Here, we introduce dynamic split computing, where the optimal split location is dynamically selected based on the state of the communication channel. By using natural bottlenecks that already exist in modern DNN architectures, dynamic split computing avoids retraining and hyperparameter optimization, and does not have any negative impact on the final accuracy of DNNs. Through extensive experiments, we show that dynamic split computing achieves faster inference in edge computing environments where the data rate and server load vary over time.
Original languageEnglish
Title of host publicationICASSP 2023 : 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publication date2023
ISBN (Electronic)978-1-7281-6327-7, 978-1-7281-6328-4
Publication statusPublished - 2023
SeriesI E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings


Dive into the research topics of 'Dynamic Split Computing for Efficient Deep Edge Intelligence'. Together they form a unique fingerprint.

Cite this