@article{ART002516912},
author={Sang Min Suh},
title={Effective Implementation for Fast Deep Learning Algorithm},
journal={Journal of Knowledge Information Technology and Systems},
issn={1975-7700},
year={2019},
volume={14},
number={5},
pages={553-561},
doi={10.34163/jkits.2019.14.5.012}
TY - JOUR
AU - Sang Min Suh
TI - Effective Implementation for Fast Deep Learning Algorithm
JO - Journal of Knowledge Information Technology and Systems
PY - 2019
VL - 14
IS - 5
PB - Korea Knowledge Information Technology Society
SP - 553
EP - 561
SN - 1975-7700
AB - AI (Artificial Intelligence) based on deep learning has been successful in many application areas. Supervised learning such as image classification and object detection has been mainly used for vision and ADAS (Advanced Driver Assistance Systems) / AD (Autonomous Driving). And reinforce learning has been generally utilized for robotics and energy optimization. Therefore, in order to improve the performance, many research papers have focused on optimizing neural networks. However, in practice, FPS (frame per second) is a hidden and critical factor because FPS is also included in the performance measurement. This note show that pre-processing and post-processing are major components affecting FPS. And It is verified that FPS cannot be improved by optimizing the neural network itself because the pre-processing and post-processing are out of the neural networks. In this note, fast pre-processing methods on the basis of DSP (digital signal processing) is suggested. For DSP implementation, binary arithmetic is presented and quantization error due to the conversion from floating point calculation to fixed point calculation is discussed. In addition, major design frameworks for deep learning algorithm implementation are compared and their merit and demerit are also summarized. In the note, implementation is categorized into three, i.e., input data generation with pre-processing, model design of neural network, and performance evaluation. With the selected framework, detailed implementation is also presented.
KW - Artificial intelligence;Deep learning;Digital signal processing;Image classification;Keras;Tensorflow
DO - 10.34163/jkits.2019.14.5.012
ER -
Sang Min Suh. (2019). Effective Implementation for Fast Deep Learning Algorithm. Journal of Knowledge Information Technology and Systems, 14(5), 553-561.
Sang Min Suh. 2019, "Effective Implementation for Fast Deep Learning Algorithm", Journal of Knowledge Information Technology and Systems, vol.14, no.5 pp.553-561. Available from: doi:10.34163/jkits.2019.14.5.012
Sang Min Suh "Effective Implementation for Fast Deep Learning Algorithm" Journal of Knowledge Information Technology and Systems 14.5 pp.553-561 (2019) : 553.
Sang Min Suh. Effective Implementation for Fast Deep Learning Algorithm. 2019; 14(5), 553-561. Available from: doi:10.34163/jkits.2019.14.5.012
Sang Min Suh. "Effective Implementation for Fast Deep Learning Algorithm" Journal of Knowledge Information Technology and Systems 14, no.5 (2019) : 553-561.doi: 10.34163/jkits.2019.14.5.012
Sang Min Suh. Effective Implementation for Fast Deep Learning Algorithm. Journal of Knowledge Information Technology and Systems, 14(5), 553-561. doi: 10.34163/jkits.2019.14.5.012
Sang Min Suh. Effective Implementation for Fast Deep Learning Algorithm. Journal of Knowledge Information Technology and Systems. 2019; 14(5) 553-561. doi: 10.34163/jkits.2019.14.5.012
Sang Min Suh. Effective Implementation for Fast Deep Learning Algorithm. 2019; 14(5), 553-561. Available from: doi:10.34163/jkits.2019.14.5.012
Sang Min Suh. "Effective Implementation for Fast Deep Learning Algorithm" Journal of Knowledge Information Technology and Systems 14, no.5 (2019) : 553-561.doi: 10.34163/jkits.2019.14.5.012
