Turning Channel Noise into an Accelerator for Over-the-Air Principal Component Analysis

Zhang，Zezhong1,2; Wang，Rui2; Li，Tengfei3; Sun，Lixin4; Lau，Vincent K.N.5; Huang，Kaibin1

10.1109/SPAWC51858.2021.9593170

2021

1948-3244

1558-2248

978-1-6654-2852-1

IEEE Workshop on Signal Processing Advances in Wireless Communications, SPAWC

2021-September

99

601-605

27-30 Sept. 2021

Lucca, Italy

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC

T In recent years, the attempts on distilling mobile data into useful knowledge have led to the deployment of machine learning algorithms at the network edge. Principal component analysis (PCA) is a classic technique for extracting the linear structure of a dataset, which is useful for feature extraction and data compression. In this work, we propose the deployment of distributed PCA over a multi-access channel based on the algorithm of stochastic gradient descent to learn the dominant feature space of a distributed dataset at multiple devices. Over- the-air aggregation is adopted to reduce the multi-access latency, giving the name over-the-air PCA. The novelty of this design lies in exploiting channel noise to accelerate the descent in the region around each saddle point encountered by gradient descent, thereby increasing the convergence speed of over-the-air PCA. The idea is materialized by proposing a power-control scheme controlling the level of channel noise accordingly. The scheme is proved to achieve faster convergence than in the case without power control by experiments on real datasets.

Wireless communication Knowledge engineering Machine learning algorithms Power control Signal processing algorithms Signal processing Feature extraction

通讯

英语

SCI ; EI

Fellowship Award from the Research Grants Council of the Hong Kong Special Administrative Region, China[HKU RFS2122-7S04] ; Guangdong Basic and Applied Basic Research Foundation[2019B1515130003] ; Hong Kong Research Grants Council["17208319","17209917"] ; Innovation and Technology Fund[GHP/016/18GD] ; Shenzhen Science and Technology Program[JCYJ20200109141414409] ; National Natural Science Foundation of China["62001310","62171213"] ; National Key Research and Development Program of China[2018YFB1800800]

Engineering ; Telecommunications

Engineering, Electrical & Electronic ; Telecommunications

WOS:000866499900010

20220311473818

Gradient methods ; Learning algorithms ; Machine learning ; Power control ; Spurious signal noise ; Stochastic systems

Machine Learning:723.4.2 ; Control Systems:731.1 ; Specific Variables Control:731.3 ; Numerical Methods:921.6 ; Mathematical Statistics:922.2 ; Systems Science:961

2-s2.0-85122795323

COMPUTER SCIENCE

Scopus

1.The University Of Hong Kong,Department Of Electrical And Electronic Engineering,Hong Kong
2.Southern University Of Science And Technology,Department Of Electrical And Electronic Engineering,China
3.Shenzhen Institute Of Radio Testing and Tech.,Shenzhen,China
4.Baicells Technologies Co.,Ltd.,United States
5.Hong Kong University Of Science And Technology,Department Of Electronic And Computer Engineering,Hong Kong

Zhang，Zezhong,Wang，Rui,Li，Tengfei,et al. Turning Channel Noise into an Accelerator for Over-the-Air Principal Component Analysis[C]:IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC,2021:601-605.