Noise vs Randomness

In this project, we try different approaches towards finding the effect of noise on randomness and attempting to classify different backends for generating quantum randomness numbers. We improve the current classical benchmark from 54% accuracy to 58% accuracy and we also try a qml algorithm to see if having a quantum data better for qml applications. We further investigate Quantum GAN successfully and compare classes pairwise. We are able to get a generative model with almost zero loss alongside a discriminator model with exponential decay in loss over 50 epochs.