期刊
IEEE TRANSACTIONS ON BIG DATA
卷 8, 期 3, 页码 644-657出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TBDATA.2018.2880978
关键词
Neural networks; Cloud computing; Ubiquitous computing; Task analysis; Computational modeling; Big Data; heterogeneous distributed deep neural network; HDDNN; deep neural network; DNN; Internet of Things; edge computing; cloud computing
资金
- National NSF of China [61525204, 61732010, 61872234]
The paper proposes a heterogeneous distributed deep neural network (HDDNN) framework for ubiquitous intelligent computing. It optimizes the utilization of hierarchical distributed systems for DNN and tailors DNN for real-world distributed systems, resulting in low response time, high performance, and better user experience.
For the pursuit of ubiquitous computing, distributed computing systems containing the cloud, edge devices, and Internet-of-Things devices are highly demanded. However, existing distributed frameworks do not tailor for the fast development of Deep Neural Network (DNN), which is the key technique behind many intelligent applications nowadays. Based on prior exploration on distributed deep neural networks (DDNN), we propose Heterogeneous Distributed Deep Neural Network (HDDNN) over the distributed hierarchy, targeting at ubiquitous intelligent computing. While being able to support basic functionalities of DNNs, our framework is optimized for various types of heterogeneity, including heterogeneous computing nodes, heterogeneous neural networks, and heterogeneous system tasks. Besides, our framework features parallel computing, privacy protection and robustness, with other consideration for the combination of heterogeneous distributed system and DNN. Extensive experiments demonstrate that our framework is capable of utilizing hierarchical distributed system better for DNN and tailoring DNN for real-world distributed system properly, which is with low response time, high performance, and better user experience.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据