Efficient quantum state tomography with mode-assisted training

Neural networks (NNs) representing quantum states are typically trained using Markov chain Monte Carlo based methods. However, unless specifically designed, such samplers only consist of local moves, making the slow-mixing problem prominent even for extremely simple states. Here, we propose to use mode-assisted training that provides global information via modes NN distribution. Applied state tomography restricted Boltzmann machines, this method improves quality reconstructed by orders magnitude. The is applicable other types NNs and may efficiently tackle problems previously unmanageable.