Taguchi Alexander T, Evans Ethan D, Dikanov Sergei A, Griffin Robert G
Department of Chemistry , Massachusetts Institute of Technology , 77 Massachusetts Avenue , Cambridge , Massachusetts 02139 , United States.
Department of Veterinary Clinical Medicine , University of Illinois at Urbana-Champaign , Urbana , Illinois 61801 , United States.
J Phys Chem Lett. 2019 Mar 7;10(5):1115-1119. doi: 10.1021/acs.jpclett.8b03797. Epub 2019 Feb 26.
A machine learning approach is presented for analyzing complex two-dimensional hyperfine sublevel correlation electron paramagnetic resonance (HYSCORE EPR) spectra with the proficiency of an expert spectroscopist. The computer vision algorithm requires no training on experimental data; rather, all of the spin physics required to interpret the spectra are learned from simulations alone. This approach is therefore applicable even when insufficient experimental data exist to train the algorithm. The neural network is demonstrated to be capable of utilizing the full information content of two-dimensional N HYSCORE spectra to predict the magnetic coupling parameters and their underlying probability distributions that were previously inaccessible. The predicted hyperfine ( a, T) and N quadrupole ( K, η) coupling constants deviate from the previous manual analyses of the experimental spectra on average by 0.11 MHz, 0.09 MHz, 0.19 MHz, and 0.09, respectively.
提出了一种机器学习方法，可像专业光谱学家一样熟练地分析复杂的二维超精细子能级相关电子顺磁共振（HYSCORE EPR）光谱。该计算机视觉算法无需对实验数据进行训练；相反，解释光谱所需的所有自旋物理知识仅从模拟中学习。因此，即使存在不足以训练该算法的实验数据，此方法也适用。结果表明，神经网络能够利用二维N HYSCORE光谱的全部信息内容，来预测以前无法获得的磁耦合参数及其潜在概率分布。预测的超精细（a，T）和N四极（K，η）耦合常数与之前对实验光谱的手动分析相比，平均偏差分别为0.11 MHz、0.09 MHz、0.19 MHz和0.09。
