Abstract:X-ray absorption spectroscopy (XAS) is a powerful characterization technique for probing the local chemical environment of absorbing atoms. However, analyzing XAS data presents with significant challenges, often requiring extensive, computationally intensive simulations, as well as significant domain expertise. These limitations hinder the development of fast, robust XAS analysis pipelines that are essential in high-throughput studies and for autonomous experimentation. We address these challenges with a suite of transfer learning approaches for XAS prediction, each uniquely contributing to improved accuracy and efficiency, as demonstrated on K-edge spectra database covering eight 3d transition metals (Ti-Cu). Our framework is built upon three distinct strategies. First, we use M3GNet to derive latent representations of the local chemical environment of absorption sites as input for XAS prediction, achieving up to order-of-magnitude improvements over conventional featurization techniques. Second, we employ a hierarchical transfer learning strategy, training a universal multi-task model across elements before fine-tuning for element-specific predictions. This cascaded approach after element-wise fine-turning yields models that outperform element-specific models by up to 31\%. Third, we implement cross-fidelity transfer learning, adapting a universal model to predict spectra generated by simulation of a different fidelity with a much higher computational cost. This approach improves prediction accuracy by up to 24\% over models trained on the target fidelity alone. Our approach is extendable to XAS prediction for a broader range of elements and offers a generalizable transfer learning framework to enhance other deep-learning models in materials science.

A graph neural network and transformer based approach to XANES data analyis

Highly sensitive 2D X-ray absorption spectroscopy via physics informed machine learning

Transmission X-ray microscopy-based three-dimensional XANES imaging

A Universal Deep Learning Framework for Materials X-ray Absorption Spectra

Robust Machine Learning Inference from X-ray Absorption Near Edge Spectra through Featurization

Subspace modeling for fast and high-sensitivity X-ray chemical imaging

Machine Learning Automated Approach for Enormous Synchrotron X-Ray Diffraction Data Interpretation

Structure-Aware Sparse-View X-ray 3D Reconstruction

Si K edge X-ray spectrum absorption interpretation using Neural Network

Decoding Structure-Spectrum Relationships with Physically Organized Latent Spaces

Integrating Machine Learning Potential and X-ray Absorption Spectroscopy for Predicting the Chemical Speciation of Disordered Carbon Nitrides

Analysis of extended X-ray absorption fine structure (EXAFS) data using artificial intelligence techniques

Machine Learning Automated Analysis of Enormous Synchrotron X-ray Diffraction Datasets

Database of ab initio L-edge X-ray absorption near edge structure

Bridging the Gap between the X-ray Absorption Spectroscopy and the Computational Catalysis Communities in Heterogeneous Catalysis: A Perspective on the Current and Future Research Directions

CuXASNet: Rapid and Accurate Prediction of Copper L-edge X-Ray Absorption Spectra Using Machine Learning

CrystalX: High-accuracy Crystal Structure Analysis Using Deep Learning

Introduction to X-Ray Absorption Spectroscopy and Its Applications in Material Science

Deep neural network for x-ray photoelectron spectroscopy data analysis

Machine Learning-Assisted Analysis of Small Angle X-ray Scattering

Deep Learning-Assisted Characterization of Nanoparticle Growth Processes: Unveiling SAXS Structure Evolution