Machine learning universally maps nonlocal quantum entropy via quadratic local correlation in nonequilibrium states

Understanding the nonlocal nature of quantum states is a major hurdle in advancing large-scale quantum computation and simulation, and Hao Liao, Huang Xuanqin, Wang Ping and colleagues at Shenzhen University and Beijing Normal University have taken an important step toward overcoming this hurdle. The research team has demonstrated a powerful new method for determining quantum mutual information, a key measure of entanglement, even in complex nonequilibrium quantum systems. Their approach utilizes a multilayer perceptron to establish a direct link between quantum mutual information and easily measurable local correlations, providing a practical route for the experimental determination of entanglement in platforms currently under development. This achievement not only simplifies the characterization of complex quantum states, but also establishes a general framework for reconstructing other important nonlocal observations, paving the way for deeper insights into phenomena such as many-body localization and thermalization.

Quantum mutual information (QMI), a fundamental measure of quantum correlation, exhibits this nonlocality, but computing it poses a direct challenge for complex quantum systems. The research team has developed a method to accurately estimate QMI using only local correlations observable in the quantum state, avoiding computationally expensive global measurements. This approach reconstructs the complete quantum state from the reduced density matrix and effectively captures nonlocal entropy through locally accessible information.

The research team demonstrated that this method can accurately estimate the QMI of a variety of quantum states, including those with strong entanglements and complex correlations. Furthermore, this work establishes a direct relationship between local correlation and nonlocal entropy, providing a new way to characterize and understand fundamental properties of quantum systems. This advance enables efficient analysis of quantum states in complex scenarios, paving the way for improved quantum algorithms and simulations.

Machine learning reveals MBL entanglement structure

This work addresses a central challenge in understanding many-body localization (MBL), a phenomenon in which disorder in quantum systems impedes thermalization. Characterizing the complex tangled structures that emerge in MBL systems is difficult using traditional methods. The authors explore machine learning, specifically neural networks, to predict quantum properties such as entanglement entropy from local measurements, a major shift away from directly computing complex quantities. The authors propose a machine learning framework that learns the relationship between local correlation and nonlocal entanglement entropy and demonstrate its universal applicability.

They utilize neural networks to map local observations to entanglement entropy and generate training data from numerical simulations of disordered quantum systems. The performance of their model is rigorously evaluated by comparing its predictions to direct calculations of entanglement entropy. The model shows generalization to different system sizes and strengths of disorder, and shows that it is learning fundamental relationships rather than just memorizing data.

Machine learning accurately predicts quantum mutual information

In this study, we introduce a novel machine learning framework based on multilayer perceptrons to efficiently predict quantum mutual information (QMI) in complex quantum systems. Scientists demonstrated that QMI, a key measure of entanglement, can be accurately determined by analyzing only easily accessible second-order correlations, bypassing the need for full quantum state tomography. This method proves to be effective in capturing the dynamic behavior of QMI over both the multibody local region and the thermalized region, significantly outperforming traditional computational techniques. In particular, the trained machine learning model exhibits universality and accurately predicts QMI dynamics over a wide range of disturbance intensities and time scales, beyond those used during initial training.

This suggests that there is an underlying universal relationship between low-order correlations and nonlocal quantum properties. The approach also demonstrates robustness to realistic measurement noise, indicating practical feasibility on an experimental platform. The researchers anticipate extending this framework to higher dimensions, long-range interaction models, and open quantum systems.