High Speed Receiver Modeling Using Generative Adversarial Networks

This paper presents a generative approach to modeling a high-speed receiver with a time series input. The model is not built with domain knowledge but learned from a wide range of channel conditions and input bitstreams to generate an eye diagram. The generated eye diagrams are similar to the simulated eye diagrams for the same scenario. We also developed a neural network model to evaluate the gen...