Quantum Machine Learning With a Bosonic Kernel Stockholm University Seminar

Abstract

Kernel methods are of current interest in quantum machine learning due to similarities shared with quantum computing in how information is processed in high-dimensional feature (Hilbert) spaces. Quantum kernels are believed to offer particular advantages when they are hard to simulate classically. Kerr nonlinearity—already a known route to universal continuous variable quantum computation—is hence a strong source of nonclassicality for machine learning.

In this talk, I will begin with a brief introduction to quantum machine learning as a whole, and kernel methods, before discussing the analytical form of a kernel which exploits the analogue features of Kerr-coupled modes—and could be implemented on quantum hardware (such as superconducting circuits). Measurements, such as one would make via displaced parity operators in a physical system, effectively produce different Wigner functions for the same synthetic data set, and comprise a labelling scheme which can be used as the basis of a simple classification training task. I will discuss the performance of a classification algorithm called a support vector machine using our Kerr kernel, and compare it to the performance of a standard classical kernel.

Carolyn Wood

Postdoctoral Scientist

Carolyn Wood is a postdoctoral researcher at the University of Queensland, in Brisbane, Australia focusing on quantum machine learning and physics at the interface between quantum mechanics and general relativity.