Department of Chemistry, Duke University, Durham, North Carolina 27708, United States.
Department of Chemistry and Institute for Sustainability and Energy, Northwestern University, Evanston, Illinois 60208-3113, United States.
J Phys Chem Lett. 2022 Aug 18;13(32):7454-7461. doi: 10.1021/acs.jpclett.2c01913. Epub 2022 Aug 5.
Two-dimensional (2D) spectroscopy encodes molecular properties and dynamics into expansive spectral data sets. Translating these data into meaningful chemical insights is challenging because of the many ways chemical properties can influence the spectra. To address the task of extracting chemical information from 2D spectroscopy, we study the capacity of simple feedforward neural networks (NNs) to map simulated 2D electronic spectra to underlying physical Hamiltonians. We examined hundreds of simulated 2D spectra corresponding to monomers and dimers with varied Franck-Condon active vibrations and monomer-monomer electronic couplings. We find the NNs are able to correctly characterize most Hamiltonian parameters in this study with an accuracy above 90%. Our results demonstrate that NNs can aid in interpreting 2D spectra, leading from spectroscopic features to underlying effective Hamiltonians.
二维（2D）光谱将分子性质和动力学编码到广阔的光谱数据集。由于化学性质可以以多种方式影响光谱，因此将这些数据转化为有意义的化学见解具有挑战性。为了从 2D 光谱中提取化学信息，我们研究了简单前馈神经网络（NN）将模拟 2D 电子光谱映射到潜在物理哈密顿量的能力。我们检查了数百个对应于具有不同 Franck-Condon 活性振动和单体-单体电子耦合的单体和二聚体的模拟 2D 光谱。我们发现，在这项研究中，神经网络能够正确地描述大多数哈密顿参数，准确率超过 90%。我们的结果表明，神经网络可以帮助解释 2D 光谱，从光谱特征到潜在的有效哈密顿量。
