Neural networks are not infallible. In this post we use a model inversion attack to discover data which was input to a target model.
