Variational Autoencoders for Precoding Matrices with High Spectral Efficiency

Neural networks have been used for channel decoding, channel detection, channel evaluation, and resource management in multi-input and multi-output (MIMO) wireless communication systems. In this paper, we consider the problem of finding a precoding matrix with high spectral efficiency (SE) using a variational autoencoder. There is an optimization procedure for finding optimal precoding matrices. Our goal is to build a less time-consuming algorithm with minimal loss of quality from the optimal one. As a solution to achieve this goal, we used two types of variational autoencoders to build precoding matrices: the classical variational autoencoder (VAE) and the conditional variational autoencoder (CVAE). Both methods can be used to study a wide range of optimal precoding matrices. The VAE and CVAE methods allow restoring the distribution of the predicted value by sampling random variables from the normal distribution at the intermediate stage of calculations. The construction of precoding matrices and their distribution for the SE objective function using VAE and CVAE methods is described in the literature for the first time.