Electrical Engineering Seminar
Fully programmable quantum machines with all-to-all connectivity via Floquet engineering, and quantum neural networks
ABSTRACT This talk will cover two distinct, but related, recent works: a theoretical proposal for a network of quantum oscillators whose interactions can be completely controlled , and a theoretical and numerical study of quantum neural networks, especially in the context of quantum-optical implementations .
In the first part, I will describe how to achieve fully programmable, all-to-all couplings between a system of N oscillators by using a common bus and a phase-modulation scheme. Such a system could be used to construct quantum annealers, quantum simulators, and quantum neural networks. I will present a proposed concrete implementation using superconducting circuits, and show numerical results demonstrating the application of such a system as a quantum annealer, with an illustration of how our design can achieve dramatically improved performance for solving combinatorial-optimization problems versus the current experimental state-of-the-art quantum annealers.
In the second part, I will discuss a way to quantify the capability of a quantum neural network, namely its memory capacity (which is a concept ported from the theoretical analysis of classical neural networks), and will show how the memory capacity of quantum neural networks differs from that of classical neural networks. I will then present a concrete proposal for a quantum neural network realized in a quantum-optics setting, and show how the capacity bound we find manifests in numerical experiments.
 T. Onodera*, E. Ng*, P.L. McMahon, in preparation.
 L.G. Wright, P.L. McMahon, in preparation.
BIO Peter McMahon is an assistant professor in the School of Applied and Engineering Physics at Cornell University. He received his Ph.D. from Stanford University in 2014 for research on semiconductor quantum optics, and subsequently studied the design and construction of special-purpose optical and opto-electronic computers for solving optimization problems as a postdoctoral researcher in Applied Physics at Stanford until his appointment at Cornell in 2019.
Contact: Liliana Chavarria at 626-395-4715 email@example.com