DATA-DRIVEN PROBABILISTIC MODELING OF WIRELESS CHANNELS USING CONDITIONAL VARIATIONAL AUTO-ENCODERS

摘要

A method performed by an artificial neural network includes determining a conditional probability distribution representing a channel based on a data set of transmit and receive sequences. The method also includes determining a latent representation of the channel based on the conditional probability distribution. The method further includes performing a channel-based function based on the latent representation.