We study the performance of Machine Learning (ML) classification techniques. Leveraging theory large deviations, we provide mathematical conditions for a ML classifier to exhibit error probabilities that vanish exponentially, say $\exp (-n\,I)$</inline-...
