Randomised benchmarking for characterizing and forecasting correlated processes

Xinfang Zhang,Zhihao Wu,Gregory A. L. White,Zhongcheng Xiang,Shun Hu,Zhihui Peng,Yong Liu,Dongning Zheng,Xiang Fu,Anqi Huang,Dario Poletti,Kavan Modi,Junjie Wu,Mingtang Deng,Chu Guo

2023-12-11

Abstract:The development of fault-tolerant quantum processors relies on the ability to control noise. A particularly insidious form of noise is temporally correlated or non-Markovian noise. By combining randomized benchmarking with supervised machine learning algorithms, we develop a method to learn the details of temporally correlated noise. In particular, we can learn the time-independent evolution operator of system plus bath and this leads to (i) the ability to characterize the degree of non-Markovianity of the dynamics and (ii) the ability to predict the dynamics of the system even beyond the times we have used to train our model. We exemplify this by implementing our method on a superconducting quantum processor. Our experimental results show a drastic change between the Markovian and non-Markovian regimes for the learning accuracies.

Quantum Physics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to characterize and predict time - dependent noise (i.e., non - Markovian noise) in open quantum systems. Specifically, the authors propose a new scheme that combines Randomized Benchmarking (RB) and supervised machine - learning methods to reconstruct the process tensor, thereby enabling:

1. **Quantifying non - Markovianity**: Through the reconstructed process tensor, metrics of the system's non - Markovianity, such as memory complexity \(M_j\) and multi - time non - Markovianity \(N_j\), as well as the mutual information \(I(x, y)\) between two time points, can be calculated.
2. **Predicting system dynamics**: Using the reconstructed open - quantum - evolution (OQE) model, it is possible to predict not only the system behavior in the training data but also the system behavior at unobserved time points.

The authors conducted experimental verification on a superconducting quantum processor. By adjusting the effective coupling strength \(\gamma_{\text{eff}}\) between the system and the environment, they demonstrated the effectiveness and predictive ability of this method in different situations from weak coupling (approximately Markovian) to strong coupling (highly non - Markovian). The experimental results show that when approaching the Markovian region, high - precision predictions can be obtained using a smaller memory size \(\chi\); in the highly non - Markovian region, although the accuracy is lower, the prediction performance can be systematically improved by increasing \(\chi\).