Quantization of Deep Neural Networks for Accurate Edge Computing
| العنوان: | Quantization of Deep Neural Networks for Accurate Edge Computing |
|---|---|
| المؤلفون: | Qing Lu, Wentao Chen, Yu Hu, Jian Zhuang, Hailong Qiu, Yiyu Shi, Xiaowei Xu, Meiping Huang, Tianchen Wang, Chutong Zhang |
| سنة النشر: | 2021 |
| مصطلحات موضوعية: | FOS: Computer and information sciences, 0303 health sciences, Computer science, Quantization (signal processing), Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition, 02 engineering and technology, 03 medical and health sciences, Range (mathematics), Hardware and Architecture, 0202 electrical engineering, electronic engineering, information engineering, Deep neural networks, 020201 artificial intelligence & image processing, Electrical and Electronic Engineering, Algorithm, Software, Edge computing, 030304 developmental biology |
| الوصف: | Deep neural networks (DNNs) have demonstrated their great potential in recent years, exceeding the per-formance of human experts in a wide range of applications. Due to their large sizes, however, compressiontechniques such as weight quantization and pruning are usually applied before they can be accommodated onthe edge. It is generally believed that quantization leads to performance degradation, and plenty of existingworks have explored quantization strategies aiming at minimum accuracy loss. In this paper, we argue thatquantization, which essentially imposes regularization on weight representations, can sometimes help toimprove accuracy. We conduct comprehensive experiments on three widely used applications: fully con-nected network (FCN) for biomedical image segmentation, convolutional neural network (CNN) for imageclassification on ImageNet, and recurrent neural network (RNN) for automatic speech recognition, and experi-mental results show that quantization can improve the accuracy by 1%, 1.95%, 4.23% on the three applicationsrespectively with 3.5x-6.4x memory reduction. 11 pages, 3 figures, 10 tables, accepted by the ACM Journal on Emerging Technologies in Computing Systems (JETC) |
| اللغة: | English |
| URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_dedup___::04515d80db859d95ba13c320061c4a2f http://arxiv.org/abs/2104.12046 |
| Rights: | OPEN |
| رقم الانضمام: | edsair.doi.dedup.....04515d80db859d95ba13c320061c4a2f |
| قاعدة البيانات: | OpenAIRE |
| الوصف غير متاح. |