Dealing with noise in digital quantum simulations and variational circuits on a trapped-ion machine
ORAL
Abstract
Our quantum computer consists of a chain of 171Yb+ ions with individual Raman beam addressing and individual readout. This fully connected system can be configured to run any sequence of single- and two-qubit gates, making it in effect an arbitrarily programmable digital quantum computer. The high degree of control can be used for digital, but also for analog and hybrid quantum simulations. We also add a classical optimization layer to our quantum stack to realize variational optimization methods [1].
Noisy operations influence all quantum computing applications and in the absence of fault-tolerant encoding, different mitigation strategies are being investigated. We recently simulated the real-time dynamics of a lattice gauge theory in 1+1 dimensions, i.e., the lattice Schwinger model, and report the comparison of different error mitigation strategies for this application [2].
Quantum classical-hybrid optimization on variational circuits is seen as a promising near-term strategy applying quantum resources to combinatorial optimization problems [3]. The noisy evolution and readout of the quantum processor, as well as cost-function landscapes devoid of guiding features like gradients, pose challenges for classical optimizers. We present an efficient "fast-and-slow" optimization strategy that switches between global Bayesian and fast local optimization to avoid local minima in large parameter spaces.
[1] D. Zhu et al., Science Advances 5, 10 (2019)
[2] N. H. Nguyen et al., PRX Quantum 3, 020324 (2022)
[3] Y. Zhu et al., Quantum Sci. Technol. 8 015007 (2023)
Noisy operations influence all quantum computing applications and in the absence of fault-tolerant encoding, different mitigation strategies are being investigated. We recently simulated the real-time dynamics of a lattice gauge theory in 1+1 dimensions, i.e., the lattice Schwinger model, and report the comparison of different error mitigation strategies for this application [2].
Quantum classical-hybrid optimization on variational circuits is seen as a promising near-term strategy applying quantum resources to combinatorial optimization problems [3]. The noisy evolution and readout of the quantum processor, as well as cost-function landscapes devoid of guiding features like gradients, pose challenges for classical optimizers. We present an efficient "fast-and-slow" optimization strategy that switches between global Bayesian and fast local optimization to avoid local minima in large parameter spaces.
[1] D. Zhu et al., Science Advances 5, 10 (2019)
[2] N. H. Nguyen et al., PRX Quantum 3, 020324 (2022)
[3] Y. Zhu et al., Quantum Sci. Technol. 8 015007 (2023)
–
Presenters
-
Norbert M Linke
Duke University
Authors
-
Norbert M Linke
Duke University