Classification Logit Two-sample Testing by Neural Networks

Xiuyuan Cheng, Alexander Cloninger

The recent success of generative adversarial networks and variational learning suggests training a classifier network may work well in addressing the classical two-sample problem.…