标题： 通过改进的量子库计算增强混沌动力学的预测能力 显示英文标题

标题： Enhanced Predictive Capability for Chaotic Dynamics by Modified Quantum Reservoir Computing

摘要： 推断随时间演化的时空混沌系统（SCS）的状态对于各种应用至关重要。然而，由于非周期动力学的复杂性和获得鲁棒解的困难，这通常是一项严峻的挑战。近年来，人们越来越关注使用经典和量子机器学习方法来解决这个问题。尽管这些现有方案在相对较短的时间内对预测SCS有效，但它们无法提供超过训练时间长度的更长时间内的鲁棒解。在这里，我们提出了一种改进混沌行为预测的方法。我们的方法可以被视为一种新颖的量子储备池计算方案，它可以同时捕捉输入数据的线性和非线性特征，并在修改后的哈密顿量下演化。我们的工作为处理SCS开辟了新的途径。 显示英文摘要

摘要： Deducing the states of spatiotemporally chaotic systems (SCSs) as they evolve in time is crucial for various applications. However, it is a dramatic challenge for generally achieving so due to the complexity of non-periodic dynamics and the hardness of obtaining robust solutions. In recent, there is a growing interest in approaching the problem using both classical and quantum machine learning methods. Although effective for predicting SCSs within a relative short time, the current schemes are not capable of providing robust solutions for longer time than training time. Here, we propose an approach for advancing the prediction of chaotic behavior. Our approach can be viewed as a novel quantum reservoir computing scheme, which can simultaneously capture the linear and the nonlinear features of input data and evolve under a modified Hamiltonian. Our work paves the way for a new avenue in handling SCSs.