Machine learning for predictive estimation of qubit dynamics subject to dephasing

Abstract

Decoherence remains a major challenge in quantum computing hardware and a variety of physical-layer controls provide opportunities to mitigate the impact of this phenomenon through feedback and feedforward control. In this work, we compare a variety of machine learning algorithms derived from diverse fields for the task of state estimation (retrodiction) and forward prediction of future qubit state evolution for a single qubit subject to classical, non-Markovian dephasing. Our approaches involve the construction of a dynamical model capturing qubit dynamics via autoregressive or Fourier-type protocols using only a historical record of projective measurements. A detailed comparison of achievable prediction horizons, model robustness, and measurement-noise-filtering capabilities for Kalman Filters (KF) and Gaussian Process Regression (GPR) algorithms is provided. We demonstrate superior performance from the autoregressive KF relative to Fourier-based KF approaches and focus on the role of filter optimization in achieving suitable performance. Finally, we examine several realizations of GPR using different kernels and discover that these approaches are generally not suitable for forward prediction. We highlight the underlying failure mechanism in this application and identify ways in which the output of the algorithm may be misidentified numerical artefacts.

33 pages (with Appendices), 9 figures

References in corpus (4)

• Reconstructing propagation networks with natural diversity and identifying hidden sources
• Suppressing qubit dephasing using real-time Hamiltonian estimation
• Prediction and retrodiction for a continuously monitored superconducting qubit
• Experimental bath engineering for quantitative studies of quantum control

Cited by in corpus (14)

• Detecting and tracking drift in quantum information processors
• Learning phase transitions from dynamics
• Real-time calibration with spectator qubits
• Cross-sensor feedback stabilization of an emulated quantum spin gyroscope
• Reinforcement learning for semi-autonomous approximate quantum eigensolver
• Supervised learning for robust quantum control in composite-pulse systems
• Real-time two-axis control of a spin qubit
• Machine learning applied to quantum synchronization-assisted probing
• Learning Order Parameters from Videos of Dynamical Phases for Skyrmions with Neural Networks
• Autonomous adaptive noise characterization in quantum computers
• Physics-informed tracking of qubit fluctuations
• Surpassing spectator qubits with photonic modes and continuous measurement for Heisenberg-limited noise mitigation
• Efficient adaptive Bayesian estimation of a slowly fluctuating Overhauser field gradient
• Adaptive filtering of projective quantum measurements using discrete stochastic methods