Machine-Learning X-Ray Absorption Spectra to Quantitative Accuracy

Abstract

Simulations of excited state properties, such as spectral functions, are often computationally expensive and therefore not suitable for high-throughput modeling. As a proof of principle, we demonstrate that graph-based neural networks can be used to predict the x-ray absorption nearedge structure spectra of molecules to quantitative accuracy. Specifically, the predicted spectra reproduce nearly all prominent peaks, with 90% of the predicted peak locations within 1 eV of the ground truth. Besides its own utility in spectral analysis and structure inference, our method can be combined with structure search algorithms to enable high-throughput spectrum sampling of the vast material configuration space, which opens up new pathways to material design and discovery.

A Critical Review of Machine Learning of Energy Materials

“…Experimentally, atomic local environments can be inferred from X‐ray adsorption spectroscopy (XAS). Several works have tried to predict the local environments by combining ML models with high‐throughput computational or experimental XAS data . Timoshenko et al have used neural networks to predict the Pt nanoparticle structure from the L‐edge X‐ray absorption near‐edge spectra.…”

Enhancing the analysis of disorder in X-ray absorption spectra: application of deep neural networks to T-jump-X-ray probe experiments

“…To address this, contemporary works have explored supervised machine learning/deep learning algorithms with a view towards mapping the relationship between XANES spectra and the electronic and geometric structures of the systems that they characterise. 8,[33][34][35][36][37][38][39][40][41][42][43][44][45][46][47] For an ab initio MD-based approach like that described in the present Article, our own deep neural network (DNN; introduced in ref. 46) could be used to accelerate the prediction of the X-ray spectra for each of the ab initio MD snapshots (the bottleneck of the strategy), opening up a fast and cost-effective route to the quantitative interpretation of T-jump pump/X-ray probe experiments.…”

The Role of Structural Representation in the Performance of a Deep Neural Network for X-ray Spectroscopy

“…To address this, a number of recent works have explored supervised machine learning/deep learning algorithms with a view towards mapping the relationship between XANES spectra and the underlying electronic and geometric structure of materials [23][24][25][26][27][28][29][30][31]. In the majority of these recent works, the authors have focused on mapping the spectrum onto a property or structure, and these works have generally been system-specific or restricted to a narrow class of systems.…”