Neural-network Quantum States (NQS)的出现显著推进了波函数拟设研究,引发了轨道空间变分蒙特卡罗(VMC)探索的复兴。本工作引入三种算法增强以降低使用NQS进行VMC优化的计算需求:自适应学习率算法、约束优化和块优化。我们在cc-pVDZ基组内对H2O和N2的复杂多参考键拉伸评估了改进后的算法,并在Ahlrichs SV基组中计算了强关联铬二聚体(Cr2)的基态能量。与耦合簇理论相比,我们的结果在相对适中的CPU成本下实现了更高的精度。本工作展示了如何利用这些策略提升优化效率与鲁棒性,为更有效地优化基于大规模受限玻尔兹曼机(RBM)的NQS开辟了新路径,标志着NQS在实际量子化学应用中的重大进展。
英文摘要:
The advent of Neural-network Quantum States (NQS) has significantly advanced wave function ansatz research, sparking a resurgence in orbital space variational Monte Carlo (VMC) exploration. This work introduces three algorithmic enhancements to reduce computational demands of VMC optimization using NQS: an adaptive learning rate algorithm, constrained optimization, and block optimization. We evaluate the refined algorithm on complex multireference bond stretches of $\rm H_2O$ and $\rm N_2$ within the cc-pVDZ basis set and calculate the ground-state energy of the strongly correlated chromium dimer ($\rm Cr_2$) in the Ahlrichs SV basis set. Our results achieve superior accuracy compared to coupled cluster theory at a relatively modest CPU cost. This work demonstrates how to enhance optimization efficiency and robustness using these strategies, opening a new path to optimize large-scale Restricted Boltzmann Machine (RBM)-based NQS more effectively and marking a substantial advancement in NQS's practical quantum chemistry applications.
N-electron valence state perturbation theory based on a density matrix renormalization group reference function, with applications to the chromium dimer and poly-p-phenylene vinylene oligomer
Time-dependent N-electron valence perturbation theory with matrix product state reference wavefunctions for large active spaces and basis sets: Applications to the chromium dimer and all-trans polyenes
在早期的工作[J. Chem. Phys. 144,064102(2016)]中,我们引入了二阶N电子价态微扰理论(t-NEVPT 2)的时间依赖性公式,其(i)具有比通常的内部收缩微扰公式更低的计算标度,并且(ii)产生完全未收缩的NEVPT 2能量。在这里,我们提出了一个组合的t-NEVPT 2与矩阵产品状态(MPS)的参考波函数(t-MPS-NEVPT 2),允许计算未收缩的动态相关能量为大的活性空间和基组,使用时间相关的密度矩阵重整化群(TD-DMRG)算法。此外,我们报告了一个低尺度的MPS为基础的实施强收缩NEVPT 2(sc-MPS-NEVPT 2),避免了计算的四粒子约简密度矩阵。我们使用这些新的方法来计算的铬二聚体的解离能和研究的低-躺在激发态的全反式多烯(C $_4$ H $_6$ C $_{24}$ H $_{26}$),将动态相关的参考波函数与多达24个活性电子和轨道。