Minimal quantum state representations from denoising auto-encoders

As multi-qubit systems increase in size, the state space scales exponentially. This makes accurate state tomography increasingly challenging and places a high demand on computational resources. This problem is compounded by the addition of experimental noise in tomographic measurements. We investigate the use of supervised machine learning, in the form of modified denoising auto-encoders, to simultaneously remove experimental noise while finding minimal latent representations of the quantum state. These representations can be later decoded into more traditional state representations.