XClose

Institute of Communications and Connected Systems

Home
Menu

Accelerated Learning-Based MIMO Detection through Weighted Neural Network Design.

ICC | Mohammad A, Masouros C, Andreopoulos Y | In this paper, we introduce a framework for a systematic acceleration of deep neural network (DNN) design for MIMO detection. A monotonically non-incr...

27 July 2020

Accelerated Learning-Based MIMO Detection through Weighted Neural Network Design.

Abstract

In this paper, we introduce a framework for a systematic acceleration of deep neural network (DNN) design for MIMO detection. A monotonically non-increasing function is used to scale the values of the layer weights such that only a certain fraction of the inputs is used for feedforward computation. This enables a dynamic weight scaling across and within the network layers, and it is termed as weight-scaling neural network-based MIMO detector (WeSNet). To increase the robustness against the changes in the activation patterns and additional enhancement in the detection accuracy for the same inference complexity, we introduce trainable weight-scaling functions. Experimental results show the superiority of our proposed method over the benchmark model (DetNet) and classical approaches based on semi-definite relaxation in terms of detection accuracy and computational efficiency.

Publication Type:Conference
Authors:Mohammad A, Masouros C, Andreopoulos Y
Publisher:IEEE
Publication date:27/07/2020
Pagination1, 6
Published proceedings:

ICC

ISBN-13978-1-7281-5089-5
Name of Conference

ICC 2020 - 2020 IEEE International Conference on Communications (ICC)

Conference placeDublin, Ireland
Conference start date07/06/2020
Conference finish date11/06/2020
DOI:10.1109/ICC40277.2020.9148726
Publisher URLhttps://ieeexplore.ieee.org/xpl/conhome/9141367/proceeding

Full Text URL:

https://discovery.ucl.ac.uk/id/eprint/10094224/

Explore how UCL research is advancing the future technologies of a connected world: