摘要
异构超密集网络(H-UDN)被认为是一种通过网络密集化来维持爆炸性的移动业务需求的解决方案。通过将接入点、处理器和存储单元放置得尽可能靠近移动用户,H-UDN带来了许多优势,包括较高的频谱利用率、较高的能量利用率和低延迟。尽管如此,H-UDNs中网络实体的高密度和多样性给协同信号的处理和资源管理带来了巨大的设计挑战。该文阐述了机器学习技术在解决这些挑战方面的巨大潜力。特别地,展示了如何利用H-UDN的图形表示来设计有效的机器学习算法。
Heterogeneous ultra-dense network(H-UDN)is envisioned as a promising solution to sustain the explosive mobile traffic demand through network densification.By placing access points,processors,and storage units as close as possible to mobile users,H-UDNs bring forth a number of advantages,including high spectral efficiency,high energy efficiency,and low latency.Nonetheless,the high density and diversity of network entities in H-UDNs introduce formidable design challenges in collaborative signal processing and resource management.This article illustrates the great potential of machine learning techniques in solving these challenges.In particular,we show how to utilize graphical representations of H-UDNs to design efficient machine learning algorithms.
作者
樊聪敏
张颖珺
袁晓军
李思贤
FAN Cong-min;ZHANG Ying-jun;YUAN Xiao-jun;LI Si-xian(Department of Information Engineering,The Chinese University of Hong Kong,HongKong China 999077;Center for Intelligent Networking and Communications,University of Electronic Science and Technology of China,Chengdu 611731)
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2020年第6期826-836,共11页
Journal of University of Electronic Science and Technology of China
基金
广东省重点领域研发计划(2018B010114001)。
关键词
深度学习
图形表式
异构超密集网络
机器学习
deep learning
graphical representations
heterogeneous ultra-dense network
machine learning
