Approximation of quantum control correction scheme using deep neural networks

TL;DR

This paper demonstrates that LSTM neural networks can effectively model the relationship between ideal and drift-affected quantum control pulses, enabling correction schemes to improve system robustness.

Contribution

It introduces a neural network-based approach to approximate quantum control correction schemes, highlighting the use of LSTM networks for modeling control pulse relationships.

Findings

LSTM networks accurately model control pulse relationships.

The approach enables analysis of control robustness.

Neural networks facilitate efficient correction schemes.

Abstract

We study the functional relationship between quantum control pulses in the idealized case and the pulses in the presence of an unwanted drift. We show that a class of artificial neural networks called LSTM is able to model this functional relationship with high efficiency, and hence the correction scheme required to counterbalance the effect of the drift. Our solution allows studying the mapping from quantum control pulses to system dynamics and then analysing the robustness of the latter against local variations in the control profile.