Quantum device fine-tuning using unsupervised embedding learning

  • 2020-01-13 17:18:54
  • N. M. van Esbroeck, D. T. Lennon, H. Moon, V. Nguyen, F. Vigneau, L. C. Camenzind, L. Yu, D. M. Zumb├╝hl, G. A. D. Briggs, D. Sejdinovic, N. Ares
  • 4


Quantum devices with a large number of gate electrodes allow for precisecontrol of device parameters. This capability is hard to fully exploit due tothe complex dependence of these parameters on applied gate voltages. Weexperimentally demonstrate an algorithm capable of fine-tuning several deviceparameters at once. The algorithm acquires a measurement and assigns it a scoreusing a variational auto-encoder. Gate voltage settings are set to optimisethis score in real-time in an unsupervised fashion. We report fine-tuning timesof a double quantum dot device within approximately 40 min.


Quick Read (beta)

loading the full paper ...