Next Article in Journal
All Fiber Mach–Zehnder Interferometer Based on Intracavity Micro-Waveguide for a Magnetic Field Sensor
Previous Article in Journal
Methodology of 3D Scanning of Intangible Cultural Heritage—The Example of Lazgi Dance
Article

CitiusSynapse: A Deep Learning Framework for Embedded Systems

High Performance Embedded System SW Research Section, Artificial Intelligence Research Laboratory, Electronics and Telecommunications Research Institute (ETRI), Daejeon 34129, Korea
*
Author to whom correspondence should be addressed.
Academic Editor: Valentino Santucci
Appl. Sci. 2021, 11(23), 11570; https://doi.org/10.3390/app112311570
Received: 28 October 2021 / Revised: 26 November 2021 / Accepted: 2 December 2021 / Published: 6 December 2021
(This article belongs to the Section Computing and Artificial Intelligence)
As embedded systems, such as smartphones with limited resources, have become increasingly popular, active research has recently been conducted on performing on-device deep learning in such systems. Therefore, in this study, we propose a deep learning framework that is specialized for embedded systems with limited resources, the operation processing structure of which differs from that of standard PCs. The proposed framework supports an OpenCL-based accelerator engine for accelerator deep learning operations in various embedded systems. Moreover, the parallel processing performance of OpenCL is maximized through an OpenCL kernel that is optimized for embedded GPUs, and the structural characteristics of embedded systems, such as unified memory. Furthermore, an on-device optimizer for optimizing the performance in on-device environments, and model converters for compatibility with conventional frameworks, are provided. The results of a performance evaluation show that the proposed on-device framework outperformed conventional methods. View Full-Text
Keywords: deep learning framework; embedded systems; on-device; OpenCL acceleration deep learning framework; embedded systems; on-device; OpenCL acceleration
Show Figures

Graphical abstract

MDPI and ACS Style

Hong, S.; Cho, H.; Kim, J.-S. CitiusSynapse: A Deep Learning Framework for Embedded Systems. Appl. Sci. 2021, 11, 11570. https://doi.org/10.3390/app112311570

AMA Style

Hong S, Cho H, Kim J-S. CitiusSynapse: A Deep Learning Framework for Embedded Systems. Applied Sciences. 2021; 11(23):11570. https://doi.org/10.3390/app112311570

Chicago/Turabian Style

Hong, Seungtae, Hyunwoo Cho, and Jeong-Si Kim. 2021. "CitiusSynapse: A Deep Learning Framework for Embedded Systems" Applied Sciences 11, no. 23: 11570. https://doi.org/10.3390/app112311570

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop