Norbert Linke: Quantum Machine Learning with Trapped Ions
0A talk by Norbert Linke at the Quantum Machine Learning Workshop, hosted September 24-28, 2018 by the Joint Center for Quantum Information and Computer Science at the University of Maryland (QuICS).
Abstract: To realize the potential of quantum machine learning, a scalable hardware platform is required, for which trapped ions are a promising candidate. We present a modular quantum computing architecture comprised of a chain of 171Yb+ ions with individual Raman beam addressing and individual readout [1]. We use the transverse modes of motion in the chain to produce entangling gates between any qubit pair. This creates a fully connected system which can be configured to run any sequence of single- and two-qubit gates, making it in effect an arbitrarily programmable quantum computer. We have applied this versatile quantum system in a number of different demonstrations relating to machine learning in a quantum-classical hybrid approach, such as error mitigation in the processor itself [2], finding the ground state binding energy of the deuteron nucleus, the training of shallow circuits [3], and the preparation of quantum critical states using a quantum approximate optimization algorithm (QAOA) scheme. Recent results from these efforts, and concepts for scaling up the architecture will be discussed.
[1] S. Debnath et al., Nature 563:63 (2016)
[2] A. Seif et al., J. Phys. B 51 174006 (2018)
[3] M. Benedetti et al., arXiv:1801.07686 (2018)