Random Fourier Feature Based Deep Learning for Wireless Communications

Abstract

Deep-learning (DL) has emerged as a powerful machine-learning technique for several classic problems encountered in generic wireless communications. Specifically, random Fourier Features (RFF) based deep-learning has emerged as an attractive solution for several machinelearning problems; yet there is a lacuna of rigorous results to justify the viability of RFF based DL-algorithms in general. To address this gap, we attempt to analytically quantify the viability of RFF based DL. Precisely, in this paper, analytical proofs are presented demonstrating that RFF based DL architectures have lower approximationerror and probability of misclassification as compared to classical DL architectures. In addition, a new distributiondependent RFF is proposed to facilitate DL architectures with low training-complexity. Through computer simulations, the practical application of the presented analytical results and the proposed distribution-dependent RFF, are depicted for various machine-learning problems encountered in next-generation communication systems such as: a) line of sight (LOS)/non-line of sight (NLOS) classification, and b) message-passing based detection of low-density parity check codes (LDPC) codes over nonlinear visible light communication (VLC) channels. Especially in the low training-data regime, the presented simulations show that significant performance gains are achieved when utilizing RFF maps of observations. Lastly, in all the presented simulations, it is observed that the proposed distributiondependent RFFs significantly outperform RFFs, which make them useful for potential machine-learning/DL based applications in the context of next-generation communication systems.

Kernel Recursive Maximum Versoria Criterion Based Post-Distorter for VLC Using Kernel-Width Sampling

“…However, performance of polynomial filtering based approaches is limited by approximation errors caused by finite order truncation of polynomial series [13]. Furthermore, the existing ANN/DNN based equalizers (like multilayer perceptron, radial basis functions) are non-convex [14, Table 1.1], computationally complex, and are sensitive to choice of various hyperparameters like number of hidden layers, number of nodes in each layer, choice of activation function [15]. Alternatively, Alternatively, RKHS based approaches are convex, computationally efficient, and are found to outperform the conventional polynomial filters since the parameterization of the inverse of LED nonlinearity (or any nonlinear transformation in general) is guaranteed to exist in RKHS, and to learn this parameter, the apriori knowledge of LED characteristics at the receiver is not required [14].…”