arXiv subjects

Attila Cangi

Publications and source records attributed to Attila Cangi.

At least 19 recordsLinked to original sources

arXiv2026

MANDALA: An E(3)-Equivariant Graph Neural Network Framework for Learning Electronic-Structure Operators with Observable Guidance

Electronic-structure calculations based on Kohn-Sham density functional theory remain indispensable in computational materials science and chemistry. Their computational cost, however, limits accessible system sizes and simulation times. At the same time, conventional machine-learning interatomic potentials (MLIPs), which are becoming the workhorse of large-scale materials modeling, usually target only energies and forces. They therefore leave out the quantum-operator-level information required to reconstruct band structures, densities of states, spatial charge distributions, and other electronic observables. \texttt{Mandala} fills this methodological gap. It is a modular software framework for learning block-sparse electronic-structure matrices with E(3)-equivariant graph neural networks. The framework is built around a unified representation of atom-resolved Hamiltonian, overlap, and density matrices, together with reusable abstractions for basis conversion, sparse block handling, irreducible representation mapping, graph construction, model definition, and training. This design allows \texttt{Mandala} to support heterogeneous chemical compositions, a wide range of neural architecture variants within one workflow, and multiple electronic-structure backends. \texttt{Mandala} evaluates selected observables directly from the predicted operators, including band energy, electron count, density of states, and band structure. This connects electronic-structure learning and observable-guided modeling while retaining a representation tied to quantum-mechanical operators rather than only scalar or vector targets as in MLIPs. In this form, \texttt{Mandala} is intended to complement atomistic interatomic potential workflows by resolving electronic structure and operator-derived observables within one scalable implementation.

cond-mat.mtrl-sci↗

arXiv2026

Can MACE Potentials Accurately Describe Magnetism and Phase Stability in Fe-Ni Alloys? A Systematic Benchmark

We present a systematic benchmark of MACE potentials for iron-nickel alloys, focusing on structural, elastic, magnetic, and finite-temperature properties relevant to phase stability. The reference dataset comprises spin-polarized PBE density functional theory (DFT) calculations for chemically disordered special quasirandom structures (SQS), spanning compositions, bcc and fcc crystal structures, and volumetric and shear deformations. A system-specific MACE-sqs model trained on this dataset achieves validation errors of 2.0 meV/atom for energies and 24.3 meV/Angstrom for forces. Compared with several MACE foundation models, including models trained with Hubbard U corrections, MACE-sqs gives the most consistent agreement with DFT and experiment for equations of state, equilibrium volumes, elastic constants, and thermal expansion trends in bcc and fcc Fe-Ni alloys. For the bcc-to-hcp transition, MACE-sqs predicts a pure-Fe transition pressure closer to experiment than the tested foundation models, but all models predict an incorrect increase of transition pressure with Ni content. This failure indicates that high-pressure magnetic collapse and composition-dependent magnetoelastic effects are not yet fully captured. Overall, targeted SQS-based training substantially improves the accuracy of MACE potentials for Fe-Ni alloys, while phase stability under magnetic collapse remains a key limitation for future model development.

cond-mat.mtrl-sci↗

arXiv2026

Thermal PBE in warm dense matter: Does it matter and is it accurate?

Conditional probability density functional theory has recently been used to derive the temperature dependence of the Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA) for the exchange-correlation (XC) free energy. We implement and systematically benchmark thermal PBE within Kohn-Sham density functional theory calculations of warm dense matter. Comparisons with the local density approximation (LDA) and PBE functionals, as well as thermal LDA, show that thermal PBE significantly improves the description of warm dense matter properties, including energies, forces, pressures, and electronic charge densities. In particular, thermal PBE exhibits close agreement with path integral Monte Carlo (PIMC) reference data at negligible additional computational cost. This work demonstrates the practical utility of thermal PBE as an accurate semilocal functional for simulations in the warm dense regime.

cond-mat.mtrl-sci↗

arXiv2026

Molecules Meet Language: Confound-Aware Representation Learning and Chemical Property Steering in Transformer-VAE Latent Spaces

Molecular generative models often assume meaningful latent geometry, but apparent property predictability can reflect sequence-level shortcuts rather than chemical organization. We study this issue in an unsupervised autoregressive Transformer-VAE trained on SELFIES. After training, we freeze the model, fit linear probes to RDKit descriptors, and use the probe weights as candidate global steering directions. To separate chemical signal from SELFIES artifacts, we introduce a confound-aware evaluation based on residualization, confound-direction alignment analysis, and decoded-molecule traversal. This is necessary because SELFIES length, branch tokens, ring tokens, and token entropy are strongly encoded in the latent space. Under this confound-aware evaluation, we find robust monotonic steering for cLogP, FractionCSP3, HeavyAtomCount, TPSA, BertzCT, and HBA. Nonlinear probes further show that some properties admit stable global directions, while others are better described by local latent gradients. Overall, our results show that chemically meaningful steering can emerge in entangled molecular latent spaces, but only when validated through decoded molecules and controlled for representation-level confounds.

cs.LG↗

arXiv2025

Machine Learning Time Propagators for Time-Dependent Density Functional Theory Simulations

Time-dependent density functional theory (TDDFT) is a widely used method to investigate electron dynamics under external time-dependent perturbations such as laser fields. In this work, we present a machine learning approach to accelerate electron dynamics simulations based on real time TDDFT using autoregressive neural operators as time-propagators for the electron density. By leveraging physics-informed constraints and featurization, and high-resolution training data, our model achieves superior accuracy and computational speed compared to traditional numerical solvers. We demonstrate the effectiveness of our model on a class of one-dimensional diatomic molecules under the influence of a range of laser parameters. This method has potential in enabling on-the-fly modeling of laser-irradiated molecules and materials by utilizing fast machine learning predictions in a large space of varying experimental parameters of the laser.

cond-mat.mtrl-sci↗

arXiv2025

Prediction of atomic H adsorption energies in metalloid doped MSSe (M = Mo/W) Janus layers: A combined DFT and machine learning study

Janus derivatives of 2H MX2 (M = Mo/W; X = S/Se), namely MSSe, have already been experimentally realized and explored for applications in photocatalysis, photovoltaics, and optoelectronics. Focusing on the photocatalytic properties of these layers, we investigate the adsorption of atomic hydrogen on the MSSe layers in the presence of metalloid dopants B, Si, and Ge. The layers in their pristine form exhibit positive adsorption energies, indicating an endothermic nature. Substitution of a dopant in the pristine MSSe layers alters the local symmetry, bonding character, and charge distribution, thereby increasing the number of active sites for hosting H adsorption and reducing the adsorption energy. We select distinct sites, both atomic and interstitial, for the substitution of dopants. The energetics of the H atom at various sites is studied to find the most favorable active site on the MSSe Janus layers. Our results based on density functional theory calculations show that the adsorption process becomes spontaneous and less attractive in the presence of atomic site dopant substitution, whereas the interstitial site results in an endothermic behavior. Moreover, having the data from DFT, we develop a supervised machine learning model for predicting the hydrogen adsorption energy. For this purpose, we utilize 23 elemental features of the atoms involved in the structure, thereby eliminating the need for DFT calculations in feature design. The dimensionality reduction technique, principal component analysis, is employed to reduce the dimensionality of the feature space, yielding independent features that are mutually orthogonal. The model is implemented as a multi-layer perceptron regressor with two hidden layers. The data augmentation technique is employed to artificially expand the dataset size, thereby enhancing the accuracy of the neural network model by 0.90% on the testing data.

cond-mat.mtrl-sci↗

arXiv2025

Scalable Machine Learning Model for Energy Decomposition Analysis in Aqueous Systems

Energy decomposition analysis (EDA) based on absolutely localized molecular orbitals provides detailed insights into intermolecular bonding by decomposing the total molecular binding energy into physically meaningful components. Here, we develop a neural network EDA model capable of predicting the electron delocalization energy component of water molecules, which captures the stabilization arising from charge transfer between occupied absolutely localized molecular orbitals of one molecule and the virtual orbitals of another. Exploiting the locality assumption of the electronic structure, our model enables accurate prediction of electron delocalization energies for molecular systems far beyond the size accessible to conventional density functional theory calculations, while maintaining its accuracy. We demonstrate the applicability of our approach by modeling hydration effects in large molecular complexes, specifically in metal-organic frameworks.

physics.chem-ph↗

arXiv2025

Strong geometry dependence of the X-ray Thomson Scattering Spectrum in single crystal silicon

We report on results from an experiment at the European XFEL where we measured the x-ray Thomson scattering (XRTS) spectrum of single crystal silicon with ultrahigh resolution. Compared to similar previous experiments, we consider a more complex scattering setup, in which the scattering vector changes orientation through the crystal lattice. In doing so, we are able to observe strong geometric dependencies in the inelastic scattering spectrum of silicon at low scattering angles. Furthermore, the high quality of the experimental data allows us to benchmark state-of-the-art TDDFT calculations, and demonstrate TDDFT's ability to accurately predict these geometric dependencies. Finally, we note that this experimental data was collected at a much faster rate than another recently reported dataset using the same setup, demonstrating that ultrahigh resolution XRTS data can be collected in more general experimental scenarios.

cond-mat.mtrl-sci↗

arXiv2024

Materials Learning Algorithms (MALA): Scalable Machine Learning for Electronic Structure Calculations in Large-Scale Atomistic Simulations

We present the Materials Learning Algorithms (MALA) package, a scalable machine learning framework designed to accelerate density functional theory (DFT) calculations suitable for large-scale atomistic simulations. Using local descriptors of the atomic environment, MALA models efficiently predict key electronic observables, including local density of states, electronic density, density of states, and total energy. The package integrates data sampling, model training and scalable inference into a unified library, while ensuring compatibility with standard DFT and molecular dynamics codes. We demonstrate MALA's capabilities with examples including boron clusters, aluminum across its solid-liquid phase boundary, and predicting the electronic structure of a stacking fault in a large beryllium slab. Scaling analyses reveal MALA's computational efficiency and identify bottlenecks for future optimization. With its ability to model electronic structures at scales far beyond standard DFT, MALA is well suited for modeling complex material systems, making it a versatile tool for advanced materials research.

cond-mat.mtrl-sci↗

arXiv2024

Electrical Conductivity of Warm Dense Hydrogen from Ohm's Law and Time-Dependent Density Functional Theory

Understanding the electrical conductivity of warm dense hydrogen is critical for both fundamental physics and applications in planetary science and inertial confinement fusion. We demonstrate how to calculate the electrical conductivity using the continuum form of Ohm's law, with the current density obtained from real-time time-dependent density functional theory. This approach simulates the dynamic response of hydrogen under warm dense matter conditions, with temperatures around 30,000 K and mass densities ranging from 0.02 to 0.98 g/cc. We systematically address finite-size errors in real-time time-dependent density functional theory, demonstrating that our calculations are both numerically feasible and reliable. Our results show good agreement with other approaches, highlighting the effectiveness of this method for modeling electronic transport properties from ambient to extreme conditions.

cond-mat.mtrl-sci↗

arXiv2024

Effects of Mosaic Crystal Instrument Functions on X-ray Thomson Scattering Diagnostics

Mosaic crystals, with their high integrated reflectivities, are widely-employed in spectrometers used to diagnose high energy density systems. X-ray Thomson scattering (XRTS) has emerged as a powerful diagnostic tool of these systems, providing in principle direct access to important properties such as the temperature via detailed balance. However, the measured XRTS spectrum is broadened by the spectrometer instrument function (IF), and without careful consideration of the IF one risks misdiagnosing system conditions. Here, we consider in detail the IF of 40 $μ$m and 100 $μ$m mosaic HAPG crystals, and how the broadening varies across the spectrometer in an energy range of 6.7-8.6 keV. Notably, we find a strong asymmetry in the shape of the IF towards higher energies. As an example, we consider the effect of the asymmetry in the IF on the temperature inferred via XRTS for simulated 80 eV CH plasmas, and find that the temperature can be overestimated if an approximate symmetric IF is used. We therefore expect a detailed consideration of the full IF will have an important impact on system properties inferred via XRTS in both forward modelling and model-free approaches.

physics.plasm-ph↗

arXiv2024

Accelerating Electron Dynamics Simulations through Machine Learned Time Propagators

Time-dependent density functional theory (TDDFT) is a widely used method to investigate electron dynamics under various external perturbations such as laser fields. In this work, we present a novel approach to accelerate real time TDDFT based electron dynamics simulations using autoregressive neural operators as time-propagators for the electron density. By leveraging physics-informed constraints and high-resolution training data, our model achieves superior accuracy and computational speed compared to traditional numerical solvers. We demonstrate the effectiveness of our model on a class of one-dimensional diatomic molecules. This method has potential in enabling real-time, on-the-fly modeling of laser-irradiated molecules and materials with varying experimental parameters.

cond-mat.mtrl-sci↗

arXiv2024

Ultrahigh Resolution X-ray Thomson Scattering Measurements at the European XFEL

Using a novel ultrahigh resolution ($ΔE \sim 0.1\,$eV) setup to measure electronic features in x-ray Thomson scattering (XRTS) experiments at the European XFEL in Germany, we have studied the collective plasmon excitation in aluminium at ambient conditions, which we can measure very accurately even at low momentum transfers. As a result, we can resolve previously reported discrepancies between ab initio time-dependent density functional theory simulations and experimental observations. The demonstrated capability for high-resolution XRTS measurements will be a game changer for the diagnosis of experiments with matter under extreme densities, temperatures, and pressures, and unlock the full potential of state-of-the-art x-ray free electron laser (XFEL) facilities to study planetary interior conditions, to understand inertial confinement fusion applications, and for material science and discovery.

physics.plasm-ph↗

arXiv2023

Inverting the Kohn-Sham equations with physics-informed machine learning

Electronic structure theory calculations offer an understanding of matter at the quantum level, complementing experimental studies in materials science and chemistry. One of the most widely used methods, density functional theory (DFT), maps a set of real interacting electrons to a set of fictitious non-interacting electrons that share the same probability density. Ensuring that the density remains the same depends on the exchange-correlation (XC) energy and, by a derivative, the XC potential. Inversions provide a method to obtain exact XC potentials from target electronic densities, in hopes of gaining insights into accuracy-boosting approximations. Neural networks provide a new avenue to perform inversions by learning the mapping from density to potential. In this work, we learn this mapping using physics-informed machine learning (PIML) methods, namely physics informed neural networks (PINNs) and Fourier neural operators (FNOs). We demonstrate the capabilities of these two methods on a dataset of one-dimensional atomic and molecular models. The capabilities of each approach are discussed in conjunction with this proof-of-concept presentation. The primary finding of our investigation is that the combination of both approaches has the greatest potential for inverting the Kohn-Sham equations at scale.

physics.comp-ph↗

arXiv2023

Physics-enhanced neural networks for equation-of-state calculations

Rapid access to accurate equation-of-state (EOS) data is crucial in the warm-dense matter regime, as it is employed in various applications, such as providing input for hydrodynamic codes to model inertial confinement fusion processes. In this study, we develop neural network models for predicting the EOS based on first-principles data. The first model utilizes basic physical properties, while the second model incorporates more sophisticated physical information, using output from average-atom calculations as features. Average-atom models are often noted for providing a reasonable balance of accuracy and speed; however, our comparison of average-atom models and higher-fidelity calculations shows that more accurate models are required in the warm-dense matter regime. Both the neural network models we propose, particularly the physics-enhanced one, demonstrate significant potential as accurate and efficient methods for computing EOS data in warm-dense matter.

physics.comp-ph↗

arXiv2023

Probing Iron in Earth's Core With Molecular-Spin Dynamics

Dynamic compression of iron to Earth-core conditions is one of the few ways to gather important elastic and transport properties needed to uncover key mechanisms surrounding the geodynamo effect. Herein a new machine-learned ab-initio derived molecular-spin dynamics (MSD) methodology with explicit treatment for longitudinal spin-fluctuations is utilized to probe the dynamic phase-diagram of iron. This framework uniquely enables an accurate resolution of the phase-transition kinetics and Earth-core elastic properties, as highlighted by compressional wave velocity and adiabatic bulk moduli measurements. In addition, a unique coupling of MSD with time-dependent density functional theory enables gauging electronic transport properties, critically important for resolving geodynamo dynamics.

cond-mat.mtrl-sci↗

arXiv2023

Machine Learning-Driven Structure Prediction for Iron Hydrides

We created a computational workflow to analyze the potential energy surface (PES) of materials using machine-learned interatomic potentials in conjunction with the minima hopping algorithm. We demonstrate this method by producing a versatile machine-learned interatomic potential for iron hydride via a neural network using an iterative training process to explore its energy landscape under different pressures. To evaluate the accuracy and comprehend the intricacies of the PES, we conducted comprehensive crystal structure predictions using our neural network-based potential paired with the minima hopping approach. The predictions spanned pressures ranging from ambient to 100 GPa. Our results reproduce the experimentally verified global minimum structures such as \textit{dhcp}, \textit{hcp}, and \textit{fcc}, corroborating previous findings. Furthermore, our in-depth exploration of the iron hydride PES at different pressures has revealed complex alterations and stacking faults in these phases, leading to the identification of several new low-enthalpy structures. This investigation has not only confirmed the presence of regions of established FeH configurations but has also highlighted the efficacy of using data-driven, extensive structure prediction methods to uncover the multifaceted PES of materials.

cond-mat.mtrl-sci↗

arXiv2023

Machine learning the electronic structure of matter across temperatures

We introduce machine learning (ML) models that predict the electronic structure of materials across a wide temperature range. Our models employ neural networks and are trained on density functional theory (DFT) data. Unlike most other ML models that use DFT data, our models directly predict the local density of states (LDOS) of the electronic structure. This provides several advantages, including access to multiple observables such as the electronic density and electronic total free energy. Moreover, our models account for both the electronic and ionic temperatures independently, making them ideal for applications like laser-heating of matter. We validate the efficacy of our LDOS-based models on a metallic test system. They accurately capture energetic effects induced by variations in ionic and electronic temperatures over a broad temperature range, even when trained on a subset of these temperatures. These findings open up exciting opportunities for investigating the electronic structure of materials under both ambient and extreme conditions.

cond-mat.mtrl-sci↗