Masters defense: Felix Frohnert

Variational Autoencoder Analysis of Quantum Systems

Representations learning of quantum systems is a challenging task at the interface of machine learning and quantum physics that has attracted increasing attention in recent years. Training unsupervised machine learning models to learn features that best describe data generated in a quantum physical process without human intervention can uncover novel representations of such data and provide us with valuable insights into how computers learn to efficiently describe quantum physical systems. In this work, we examine a deep generative model of quantum states where a variational autoencoder is trained on density matrices generated by a 2-qubit quantum circuit.