Quantum Neural Networks

Neural networks based on quantum elements may lead to more efficient ways to realize neural networks. Our research is directed towards building small feed forward neural networks consisting of a few qubits that work in the quantum regime and investigate the question on whether to expect a quantum advantage.
Team
Stefan Filipp, Gerhard Huber, Federico Roy, Franz X. Haslbeck
Recent publications
Characterization and Tomography of a Hidden Qubit
M. Pechal, G. Salis, M. Ganzhorn, D. J. Egger, M. Werninghaus, and S. Filipp
Research Article | Physical Review X 11, 041032  (2021)
Preprint: arXiv:2011.08987

A specific way to realize quantum neuron networks is based on adiabatic ramp quantum neurons. These are implemented via tunable ZZ-type couplings together with an adiabatically varying driving field. Along with higher connectivity brought by multi-qubit couplers we aim to realize first feed-forward networks and explore its capabilities beyond classical approximation.