On February 1st, 2019, Lorenzo started our PhD Seminar Series with a nice introduction to Quantum Computing on the D-Wave machine.

The development of quantum-classical hybrid algorithms is critical to deploy noisy intermediate-scale quantum devices in state-of-the-art computational models. In this seminar I will introduce the architecture of the D-Wave quantum annealer and show how this device can be used to train a Quantum Variational Autoencoder (QVAE). QVAE consists of a quantum generative process realized by a quantum Boltzmann machine and a classical autoencoding structure realized by traditional deep neural networks. The hybrid structure of QVAE allows using current-generation quantum annealers in a simple model with convolutional neural networks to achieve competitive performance on common datasets such as MNIST, Omniglot and Fashion MNIST.