Swish-Driven GoogleNet for Intelligent Analog Beam Selection in Terahertz Beamspace MIMO

In this paper, we propose an intelligent analog beam selection strategy in a terahertz (THz) band beamspace multipleinput multiple-output (MIMO) system. First inspired by transfer learning, we fine-tune the pre-trained off-the-shelf GoogleNet classifier, to learn analog beam selection as a multi-class mapping problem. Simulation results show 83% accuracy for the analog beam selection, which subsequently results in 12% spectral efficiency (SE) gain, upon the existing counterparts. Towards a more accurate classifier, we replace the conventional rectified linear unit (ReLU) activation function of the GoogleNet with the recently proposed Swish and retrain the fine-tuned GoogleNet to learn analog beam selection. It is numerically indicated that the fine-tuned Swish-driven GoogleNet achieves 86% accuracy, as well as 18% improvement in achievable SE, upon the similar schemes. Eventually, a strong ensembled classifier is developed to learn analog beam selection by sequentially training multiple fine-tuned Swish-driven GoogleNet classifiers. According to the simulations, the strong ensembled model is 90% accurate and yields 27% gain in achievable SE, in comparison with prior methods.

[1]  Thomas G. Dietterich What is machine learning? , 2020, Archives of Disease in Childhood.

[2]  Ke Liu,et al.  Low Complexity Beam Selection Scheme for Terahertz Systems: A Machine Learning Approach , 2019, 2019 IEEE International Conference on Communications Workshops (ICC Workshops).

[3]  Gilles Celeux,et al.  A Component-Wise EM Algorithm for Mixtures , 2001, 1201.5913.

[4]  Desmond P. Taylor,et al.  A Statistical Model for Indoor Multipath Propagation , 2007 .

[5]  Lajos Hanzo,et al.  Beamspace Precoding and Beam Selection for Wideband Millimeter-Wave MIMO Relying on Lens Antenna Arrays , 2019, IEEE Transactions on Signal Processing.

[6]  Chen Hu,et al.  Knowledge-Aided Deep Learning for Beamspace Channel Estimation in Millimeter-Wave Massive MIMO Systems , 2019, ArXiv.

[7]  Dumitru Erhan,et al.  Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[8]  Sergey Ioffe,et al.  Rethinking the Inception Architecture for Computer Vision , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  Robert W. Heath,et al.  5G MIMO Data for Machine Learning: Application to Beam-Selection Using Deep Learning , 2018, 2018 Information Theory and Applications Workshop (ITA).

[10]  Christos Masouros,et al.  Low RF-Complexity Millimeter-Wave Beamspace-MIMO Systems by Beam Selection , 2015, IEEE Transactions on Communications.

[11]  Akbar M. Sayeed,et al.  Beamspace MIMO for Millimeter-Wave Communications: System Architecture, Modeling, Analysis, and Measurements , 2013, IEEE Transactions on Antennas and Propagation.

[12]  Lakshmipriya Balagourouchetty,et al.  GoogLeNet-Based Ensemble FCNet Classifier for Focal Liver Lesion Diagnosis , 2020, IEEE Journal of Biomedical and Health Informatics.

[13]  Hai Lin,et al.  An Overview of Enhanced Massive MIMO With Array Signal Processing Techniques , 2019, IEEE Journal of Selected Topics in Signal Processing.