TY - GEN
T1 - Approximate Acceleration for CNN-based Applications on IoT Edge Devices
AU - Castro-Godinez, Jorge
AU - Hernandez-Araya, Deykel
AU - Shafique, Muhammad
AU - Henkel, Jorg
N1 - Funding Information:
VII. ACKNOWLEDGMENT Authors thank Alexander Leiva for his support in implementing most of the design here presented. This work was supported in part by the Instituto Tecnológico de Costa Rica.
Funding Information:
Authors thank Alexander Leiva for his support in implementing most of the design here presented. This work was supported in part by the Instituto Tecnológico de Costa Rica.
Publisher Copyright:
© 2020 IEEE.
PY - 2020/2
Y1 - 2020/2
N2 - Machine learning based sub-systems are increasingly becoming part of IoT edge devices, thereby requiring resource-efficient architectures and implementations, especially when subjected to battery-constrained scenarios. The non-exact nature of Convolutional Neural Networks (CNNs) opens the possibility to use approximate computations to reduce their required runtime and energy consumption on resource-constrained IoT edge devices without significantly compromising their classification output. In this paper, we propose a resilience exploration method and a novel approximate accelerator to speed up the execution of the convolutional layer, which is the most time consuming component of CNNs, for IoT edge devices. Trained CNNs with Caffe framework are executed on a System-on-Chip with reconfigurable hardware available, where the approximate accelerator is deployed. CNN applications developed with Caffe can take advantage of our proposed approximate acceleration to execute them on IoT edge devices.
AB - Machine learning based sub-systems are increasingly becoming part of IoT edge devices, thereby requiring resource-efficient architectures and implementations, especially when subjected to battery-constrained scenarios. The non-exact nature of Convolutional Neural Networks (CNNs) opens the possibility to use approximate computations to reduce their required runtime and energy consumption on resource-constrained IoT edge devices without significantly compromising their classification output. In this paper, we propose a resilience exploration method and a novel approximate accelerator to speed up the execution of the convolutional layer, which is the most time consuming component of CNNs, for IoT edge devices. Trained CNNs with Caffe framework are executed on a System-on-Chip with reconfigurable hardware available, where the approximate accelerator is deployed. CNN applications developed with Caffe can take advantage of our proposed approximate acceleration to execute them on IoT edge devices.
KW - accelerator architectures
KW - Approximate computing
KW - convolutional neural networks
KW - edge computing
UR - http://www.scopus.com/inward/record.url?scp=85084325560&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85084325560&partnerID=8YFLogxK
U2 - 10.1109/LASCAS45839.2020.9069040
DO - 10.1109/LASCAS45839.2020.9069040
M3 - Conference contribution
AN - SCOPUS:85084325560
T3 - 2020 IEEE 11th Latin American Symposium on Circuits and Systems, LASCAS 2020
BT - 2020 IEEE 11th Latin American Symposium on Circuits and Systems, LASCAS 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 11th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2020
Y2 - 25 February 2020 through 28 February 2020
ER -