Description
Non-Markovian dynamics is a practically relevant feature of open quantum systems and can significantly influence the behaviour of noisy quantum devices. However, its characterization typically requires full multi-time process tomography, which is experimentally and computationally expensive, and, therefore, difficult to use as a routine diagnostic tool. In this work, we develop a resource efficient neural network based machine learning algorithm for estimating the amount of non-Markovianity in informationally incomplete regime. In particular, our model achieves $0.98$ $R^2$ for processes obtained experimentally from IBM quantum platform. In doing so, we also provide a hardware aware experimental scheme to perform a large number of multi-time process tomography in parallel within a realistic timeframe.
| I am the presenting author | Yes |
|---|