简单梳理一下论文中的想法(二)

  最近看了一篇在IoT网络中,利用IRS去集成AirComp(Over-the-air computation 空中计算)和EB(energy beamforming 能量波束形成)技术解决Fast wireless data aggregation(WDA 快速无线数据聚合)和efficient battery recharging(高效的给电池充电)两大问题的论文:

  一、背景(background)--related words:

  对于AirComp来说,最近的一些研究如下:

  (1)有些研究发现采用模拟调制信号的并发传输更有助于无线通信中的快速函数计算;

  (2)扩展AirComp计算多种目标函数

  (3)在无线传感器网络中,利用模拟AirComp实现robust estimations of computing functions

  (4)a unform-forcing transceiver design to achieve reliable Air Comp

  (5)exploited AirComp to achieve efficient multimodal sensing and multiple functions computation

  (6)latency reduction in IoT networks without relying on the channel state information(CSI)

  (7)AirComp matches well with the model aggregation process in federated learning,so leveraged AirComp to accelerate the convergence rate of federated learning over wireless networks.

  以上都是对analog AirComp(模拟 AirComp)的研究,接下来还有对digital AirComp(数字 AirComp)的研究。

  (1)one-bit digital AirComp scheme to facilitate communication-efficient federated learning in wideband wireless networks

  (2)a digital AirComp scheme to achieve the reliable computation of nomographic functions

  但是,数字AirComp会比模拟AirComp有着更加严重的信号失真。

  对于EB

  

上一篇:分布式机器学习


下一篇:「李宏毅机器学习」机器学习介绍