XANE(3): An E(3)-Equivariant Graph Neural Network for Accurate Prediction of XANES Spectra from Atomic Structures

Vitor F. Grizzi; Luke N. Pretzie; Jiayi Xu; Cong Liu

cs.LGApr 13, 2026

XANE(3): An E(3)-Equivariant Graph Neural Network for Accurate Prediction of XANES Spectra from Atomic Structures

Authors: Vitor F. Grizzi, Luke N. Pretzie, Jiayi Xu, Cong Liu

Organizations: Chemical Sciences and Engineering Division, Argonne National Laboratory, Lemont, IL 60439, USA

Abstract

We present XANE(3), a physics-based E(3)-equivariant graph neural network for predicting X-ray absorption near-edge structure (XANES) spectra directly from atomic structures. The model combines tensor-product message passing with spherical harmonic edge features, absorber-query attention pooling, custom equivariant layer normalization, adaptive gated residual connections, and a spectral readout based on a multi-scale Gaussian basis with an optional sigmoidal background term. To improve line-shape fidelity, training is performed with a composite objective that includes pointwise spectral reconstruction together with first- and second-derivative matching terms. We evaluate the model on a dataset of 5,941 FDMNES simulations of iron oxide surface facets and obtain a spectrum mean squared error of $1.0 \times 10^{-3}$ on the test set. The model accurately reproduces the main edge structure, relative peak intensities, pre-edge features, and post-edge oscillations. Ablation studies show that the derivative-aware objective, custom equivariant normalization, absorber-conditioned attention pooling, adaptive gated residual mixing, and global background term each improve performance. Interestingly, a capacity-matched scalar-only variant achieves comparable pointwise reconstruction error but reduced derivative-level fidelity, indicating that explicit tensorial channels are not strictly required for low intensity error on this dataset, although they remain beneficial for capturing finer spectral structure. These results establish XANE(3) as an accurate and efficient surrogate for XANES simulation and offer a promising route toward accelerated spectral prediction, ML-assisted spectroscopy, and data-driven materials discovery.

Figures & tables

Figure 1: Schematic of XANE(3). The model maps an input atomic structure to a predicted XANES spectrum through an $E(3)$ -equivariant message-passing backbone, followed by absorber-conditioned attention pooling and invariant feature extraction. The fused absorber and context features are then used to predict both the coefficients of a Gaussian-plus-sigmoid spectral basis and the auxiliary edge energy $E_{0}$ .

Figure 2: Comparison between predicted and computed XANES spectra for representative test samples. The energy axis is referenced to the Fe Fermi energy, $E_{0}=7112.15$ eV. The model achieves excellent agreement with the ground-truth spectra, accurately reproducing not only the main peak structure but also finer features such as the pre-edge peak near $E=0$ and the post-edge oscillations.

Model	Total MSE	Spectrum MSE	Gradient MSE	Curvature MSE
baseline	2.3	1.0	0.6	0.6
no derivative loss	N/A	1.2	N/A	N/A
no global background	2.6	1.0	0.8	0.8
no gated residual	2.5	1.0	0.7	0.8
no attention pooling	2.6	0.9	0.8	0.8
no layer norm	5.5	2.0	2.0	1.5

Table 1: Ablation study of XANE(3). Lower is better for MSE. All columns refer to the test MSE of the relevant term in the loss function in (

\times 10^{-3}

) units.

Figure 3: Examples showing that lower pointwise spectrum MSE does not necessarily imply better spectral agreement. While panels a) and b) have smaller MSE values than panels c) and d), the latter better capture the main peak position, relative intensities, and post-edge oscillatory structure of the ground-truth spectrum.

Figure 4: Comparison between predicted and computed XANES for the version of the model without the sigmoid background function. The absence of the background introduces non-physical artifacts in the form of small oscillations at the high energy boundary above $E=7200$ eV.

Materials Property Prediction

Materials property prediction uses data and physical models to estimate how composition, structure, and processing determine a material’s behavior.

See all 174 papers

ChemGraph-XANES: An Agentic Framework for XANES Simulation and Curation
Apr 17, 2026Vitor F. Grizzi, Thang Duc Pham, Luke N. Pretzie +3Scientific Workflow Automation Computational Materials Science
Chemical Sciences and Engineering Division Argonne National Laboratory, Lemont, IL 60439, USA · Computational Science Division Argonne National Laboratory, Lemont, IL 60439, USA
Equivariant Graph Neural Networks Improve Optical Spectra Prediction for Materials Screening
Jun 17, 2026Kasper Helverskov Petersen, François R J Cornet, Martin Ovesen +3Materials Property Prediction Equivariant GNNs
Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kongens Lyngby, Denmark · Department of Physics, Technical University of Denmark, Kongens Lyngby, Denmark
Deep Scientific Reasoning under Physical Constraints: Structure-Aware Spectrum Prediction
Oct 25, 2026Yingheng Wang, Tao Yu, Shufeng Kong +3Materials Property Prediction Physics-Informed ML
Department of Computer Science, Cornell University · Griffith University · Institute of Condensed Matter and Nanosciences, Universit´e catholique de Louvain +3

Apr 17, 2026J/K move · Enter open · S save

ChemGraph-XANES: An Agentic Framework for XANES Simulation and Curation

Vitor F. Grizzi, Thang Duc Pham, Luke N. Pretzie, Jiayi Xu+2

Chemical Sciences and Engineering Division Argonne National Laboratory, Lemont, IL 60439, USA · Computational Science Division Argonne National Laboratory, Lemont, IL 60439, USA

Scientific Workflow Automation Computational Materials Science

Open paper AskPDF arXiv

Computational X-ray absorption near-edge structure (XANES) is widely used to interpret local coordination environments, oxidation states, and electronic structure in chemically complex systems. In practice, routine computational XANES at scale is often constrained by workflow complexity rather than by the simulation method. We present ChemGraph-XANES, a large-language-model (LLM)-based agentic framework for XANES simulation and analysis that combines retrieval-augmented generation (RAG)-assisted parameter selection from documentation, schema-constrained tool execution, deterministic FDMNES input generation, and provenance-aware data curation. The framework supports both direct scripted execution and natural-language orchestration, with both modes routed through a deterministic backend for structure handling, absorber and edge specification, input generation, execution, spectral extraction, and post-processing. We demonstrate three proof-of-capability use cases: RAG-assisted selection and propagation of FDMNES input parameters, structure-file-based execution, and chemistry-level natural-language specification of absorber and composition requests. In a recorded trace, a simulation parameter is retrieved from the FDMNES manual by the RAG-enabled agent and propagated into a schema-validated tool call, illustrating traceable parameter selection. We further show that the same execution pathway supports both explicit local structures and chemistry-level user inputs. Because XANES calculations are independent once inputs are defined, ChemGraph-XANES is designed to support task-parallel execution and the creation of structure-linked XANES collections. ChemGraph-XANES therefore serves as a practical agentic framework for computational spectroscopy and data generation, emphasizing constrained orchestration, reproducibility, and traceable outputs.

Abstract

Materials Property Prediction

Explore similar work

ChemGraph-XANES: An Agentic Framework for XANES Simulation and Curation

Equivariant Graph Neural Networks Improve Optical Spectra Prediction for Materials Screening

Deep Scientific Reasoning under Physical Constraints: Structure-Aware Spectrum Prediction