QuICS Special Seminar
Calibrating and re-calibrating quantum processors to achieve and maintain error rates below fault tolerance thresholds in the presence of drift will be a key engineering challenge in emerging large scale quantum information processing systems. To address this problem, we are adapting a canonical technique in classical electrical and control engineering: combining Kalman filters with linear-quadratic regulators for streaming adaptive control. We start by identifying a gate set error mode for a device. Gate set tomography (GST) provides a set of tomographic experiments that yield information about the model parameters, and techniques like robust phase estimation (RPE) allow one to extract only relevant parameters needed for
calibration. We then construct a Kalman filter based on RPE-like observations, which provides a streaming estimate of the drifting state of a device and its uncertainty. Drift is then corrected with a properly tuned linear quadratic regulator. Riccati equations play a key role in the analysis. Both Kalman filters and linear-quadratic regulators are extremely efficient protocols that can be embedded directly on classical control hardware. Our methods pave the way to autonomous calibration and regulation of large-scale quantum processors.
*We strongly encourage attendees to use their full name (and if possible, their UMD credentials) to join the zoom session.*