Title: Quantum Generative Adversarial Networks for Learning and Loading Random Distributions

Abstract: Quantum algorithms have the potential to outperform their classical counterparts in a variety of tasks. The realization of the advantage often requires the ability to load classical data efficiently into quantum states. However, the best known methods for loading generic data into an $n$-qubit state require $\mathcal{O}\left(2^n\right)$ gates. This scaling can easily predominate the complexity of a quantum algorithm and, thereby, impair potential quantum advantage. Our work demonstrates that quantum Generative Adversarial Networks (qGANs) facilitate efficient loading of generic probability distributions -- implicitly given by data samples -- into quantum states. More specifically, the qGAN scheme employs the interplay of a quantum channel, a variational quantum circuit, and a classical neural network to learn the probability distribution underlying training data samples and load it into the quantum channel. Effectively, the scheme results in a quantum channel that loads the learned distribution with $\mathcal{O}\left(poly\left(n\right)\right)$ gates. This distribution loading method can, thus, enable the exploitation of quantum advantage induced by other quantum algorithms, such as Quantum Amplitude Estimation. We implement the qGAN distribution learning and loading method with Qiskit and test it using a quantum simulation as well as actual quantum processors provided by the IBM Q Experience. Furthermore, we employ quantum simulation to demonstrate the use of the trained quantum channel in a quantum finance application.