论文标题
长期耗散量子动力学的卷积神经网络
Convolutional neural networks for long-time dissipative quantum dynamics
论文作者
论文摘要
开放量子系统动态的确切数值模拟通常需要巨大的计算资源。我们证明，由卷积层组成的深人造神经网络是预测开放量子系统的长期动态的强大工具，只要已知该系统的短时动态。在这项工作中开发的神经网络模型可以有效地模拟长期动力学，并且非常准确地在不同的动力学方面，从弱阻尼的连贯运动到不连贯的放松。该模型对与光合激发能传递有关的数据集进行了培训，并可以部署以研究在轻度收获复合物中观察到的持久量子相干现象。此外，我们的模型在与培训中使用的情况不同的初始条件下的性能很好。我们的方法大大减少了长期模拟所需的计算资源，并具有成为开放量子系统研究的宝贵工具的希望。
Exact numerical simulations of dynamics of open quantum systems often require immense computational resources. We demonstrate that a deep artificial neural network comprised of convolutional layers is a powerful tool for predicting long-time dynamics of an open quantum system provided the preceding short-time dynamics of the system is known. The neural network model developed in this work simulates long-time dynamics efficiently and very accurately across different dynamical regimes from weakly damped coherent motion to incoherent relaxation. The model was trained on a data set relevant to photosynthetic excitation energy transfer and can be deployed to study long-lasting quantum coherence phenomena observed in light-harvesting complexes. Furthermore, our model performs well for the initial conditions different than those used in the training. Our approach considerably reduces the required computational resources for long-time simulations and holds promise for becoming a valuable tool in the study of open quantum systems.
