Machine learning (ML) starts to be widely used to enhance the performance of multi-user multiple-input multiple-output (MU-MIMO) receivers. However, it is still unclear if such methods are truly competitive with respect to conventional methods in realistic scenarios and under practical constraints. In addition to enabling accurate signal reconstruction on realistic channel models, MU-MIMO receive algorithms must allow for easy adaptation to a varying number of users without the need for retraining. In contrast to existing work, we propose an machine learning (ML)-enhanced MU-MIMO receiver that builds on top of a conventional linear minimum mean squared error (LMMSE) architecture. It preserves the interpretability and scalability of the LMMSE receiver, while improving its accuracy in two ways. First, convolutional neural networks (CNNs) are used to compute an approximation of the second-order statistics of the channel estimation error which are required for accurate equalization. Second, a CNN-based demapper jointly processes a large number of orthogonal frequency-division multiplexing (OFDM) symbols and subcarriers, which allows it to compute better log likelihood ratios (LLRs) by compensating for channel aging. The resulting architecture can be used in the up- and downlink and is trained in an end-to-end manner, removing the need for hard-to-get perfect channel state information (CSI) during the training phase. Simulation results demonstrate consistent performance improvements over the baseline which are especially pronounced in high mobility scenarios.

End-to-end learning of communication systems enables joint optimization of transmitter and receiver, implemented as deep neural network (NN)-based autoencoders, over any type of channel and for an arbitrary performance metric. Recently, an alternating training procedure was proposed which eliminates the need for an explicit channel model. However, this approach requires feedback of real-valued losses from the receiver to the transmitter during training. In this paper, we first show that alternating training works even with a noisy feedback channel. Then, we design a system that learns to transmit real numbers over an unknown channel without a preexisting feedback link. Once trained, this feedback system can be used to communicate losses during alternating training of autoencoders. Evaluations over additive white Gaussian noise (AWGN) and Rayleigh block-fading (RBF) channels show that end-to-end communication systems trained using the proposed feedback system achieve the same performance as when trained with a perfect feedback link.