Deep Reinforcement Learning algorithm for joint optimization of transmitter and receiver
This repository contains the code for implementing the alternating training procedure for autoencoders described in the paper "Deep Reinforcement Learning Autoencoder with Noisy Feedback", available at: https://arxiv.org/pdf/1810.05419.pdf
The code is not fully coherent to the paper. In particular, the transmitter loss function should be implemented (cross-entropy instead of mse)