ML Assisted Superconducting Qubit Readout

Abstract

In this work, we use a new approach based on neural networks. A neural network classifier can efficiently converge to the optimal joint filter with sufficient training. To discriminate qubit states, the neural network is trained in a supervised fashion with labeled, unfiltered measurement data recorded over several microseconds for each qubit configuration. In this work, the neural networks will be evaluated based on the following criteria: i. Classification inaccuracy ii. Classification inaccuracy with increasing number of qubits iii. Required measurement time to achieve optimal classification inaccuracy iv. Difference between training and test classification inaccuracy v. Extensiveness of neural network training to achieve optimal classification inaccuracy vi. Classification inaccuracy fluctuations due to changing experimental conditions This work could potentially lead to general, neuromorphic-based techniques for readout of quantum systems and, as such, may have great impact practical implementations of quantum information processors.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 08, 2020
Source ID
HR00112010001

Entities

People

  • William D Oliver

Organizations

  • Defense Advanced Research Projects Agency
  • Massachusetts Institute of Technology

Tags

Fields of Study

  • Physics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Neural Network Machine Learning.
  • Quantum spin resonance or Electron Paramagnetic Resonance spectroscopy.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks
  • Quantum Computing