A5019215236- Machine Learning in Materials Science
- Cardiovascular Function and Risk Factors
- Computational Physics and Python Applications
- Medical Imaging Techniques and Applications
- Radiomics and Machine Learning in Medical Imaging
- Health Systems, Economic Evaluations, Quality of Life
- X-ray Diffraction in Crystallography
- Particle physics theoretical and experimental studies
- Particle Accelerators and Free-Electron Lasers
- Magnetic Properties and Applications
- Cardiac Imaging and Diagnostics
- Liver Disease Diagnosis and Treatment
- Statistical Methods in Clinical Trials
- Colorectal Cancer Screening and Detection
- Electronic and Structural Properties of Oxides
- Computational Drug Discovery Methods
- Atomic and Subatomic Physics Research
- 2D Materials and Applications
- Boron and Carbon Nanomaterials Research
- Inorganic Chemistry and Materials
- Superconducting Materials and Applications
- Advanced Chemical Physics Studies
- Metal and Thin Film Mechanics
- Nonlinear Photonic Systems
- Chalcogenide Semiconductor Thin Films
Material Measurement Laboratory
2016-2025
National Institute of Standards and Technology
2016-2025
National Institute of Standards
2018-2025
Theiss Research
2021-2023
Physical Measurement Laboratory
2022
Oak Ridge National Laboratory
2019-2022
Indian Army
2016-2021
Northwestern University
2019
SLAC National Accelerator Laboratory
2019
University of Florida
2011-2018
Deep learning (DL) is one of the fastest growing topics in materials data science, with rapidly emerging applications spanning atomistic, image-based, spectral, and textual modalities. DL allows analysis unstructured automated identification features. Recent development large databases has fueled application methods atomistic prediction particular. In contrast, advances image spectral have largely leveraged synthetic enabled by high quality forward models as well generative unsupervised...
Abstract Graph neural networks (GNN) have been shown to provide substantial performance improvements for atomistic material representation and modeling compared with descriptor-based machine learning models. While most existing GNN models predictions are based on atomic distance information, they do not explicitly incorporate bond angles, which critical distinguishing many structures. Furthermore, properties known be sensitive slight changes in angles. We present an Atomistic Line Neural...
The current predictive modeling techniques applied to Density Functional Theory (DFT) computations have helped accelerate the process of materials discovery by providing significantly faster methods scan candidates, thereby reducing search space for future DFT and experiments. However, in addition prediction error against DFT-computed properties, such models also inherit DFT-computation discrepancies experimentally measured properties. To address this challenge, we demonstrate that using...
We introduce a simple criterion to identify two-dimensional (2D) materials based on the comparison between experimental lattice constants and mainly obtained from Materials-Project (MP) density functional theory (DFT) calculation repository. Specifically, if relative difference two for specific material is greater than or equal 5%, we predict them be good candidates 2D materials. have predicted at least 1356 such For all systems satisfying our criterion, manually create single layer...
We study the crystal symmetry of few-layer 1T' MoTe2 using polarization dependence second harmonic generation (SHG) and Raman scattering. Bulk is known to be inversion symmetric; however, we find that broken for finite crystals with even numbers layers, resulting in strong SHG comparable other transition-metal dichalcogenides. Group theory analysis signals allows definitive assignment all modes clears up a discrepancy literature. The results were also compared density functional simulations...
We report the findings of a hackathon focused on exploring diverse applications large language models in molecular and materials science.
Classical force fields (FFs) based on machine learning (ML) methods show great potential for large scale simulations of solids.
Abstract Recent advances in machine learning (ML) have led to substantial performance improvement material database benchmarks, but an excellent benchmark score may not imply good generalization performance. Here we show that ML models trained on Materials Project 2018 can severely degraded new compounds 2021 due the distribution shift. We discuss how foresee issue with a few simple tools. Firstly, uniform manifold approximation and projection (UMAP) be used investigate relation between...
Extensive efforts to gather materials data have largely overlooked potential redundancy. In this study, we present evidence of a significant degree redundancy across multiple large datasets for various material properties, by revealing that up 95% can be safely removed from machine learning training with little impact on in-distribution prediction performance. The redundant is related over-represented types and does not mitigate the severe performance degradation out-of-distribution samples....
Abstract Modern data mining methods have demonstrated effectiveness in comprehending and predicting materials properties. An essential component the process of discovery is to know which material(s) will possess desirable For many properties, performing experiments density functional theory computations are costly time-consuming. Hence, it challenging build accurate predictive models for such properties using conventional due small amount available data. Here we present a framework property...
Abstract Scientific machine learning (ML) aims to develop generalizable models, yet assessments of generalizability often rely on heuristics. Here, we demonstrate in the materials science setting that heuristic evaluations lead biased conclusions ML and benefits neural scaling, through out-of-distribution (OOD) tasks involving unseen chemistry or structural symmetries. Surprisingly, many good performance across including boosted trees. However, analysis representation space shows most test...
Solar energy plays an important role in solving serious environmental problems and meeting the high demand. However, lack of suitable materials hinders further progress this technology. Here, we present largest inorganic solar cell material search till date using density functional theory (DFT) machine-learning approaches. We calculated spectroscopic limited maximum efficiency (SLME) Tran–Blaha-modified Becke–Johnson potential for 5097 nonmetallic identified 1997 candidates with SLME higher...
In this work we present a high-throughput first-principles study of elastic properties bulk and monolayer materials mainly using the vdW-DF-optB88 functional. We discuss trends on response with respect to changes in dimensionality. identify relation between exfoliation energy constants for layered that can help guide search vdW bonding materials. also predicted few novel auxetic behavior. The uncertainty structural due inclusion interactions is discussed. investigated 11 067 257 Lastly,...
A Materials Project based open-source Python tool, MPInterfaces, has been developed to automate the high-throughput computational screening and study of interfacial systems. The framework encompasses creation manipulation interface structures for solid/solid hetero-structures, solid/implicit solvents systems, nanoparticle/ligands systems; simple system-agnostic workflows in depth analysis using density-functional theory or empirical energy models. package leverages existing tools extends...
Abstract Artificial intelligence (AI) and machine learning (ML) have been increasingly used in materials science to build predictive models accelerate discovery. For selected properties, availability of large databases has also facilitated application deep (DL) transfer (TL). However, unavailability datasets for a majority properties prohibits widespread DL/TL. We present cross-property deep-transfer-learning framework that leverages trained on small different properties. test the proposed...
We present a complete set of chemo-structural descriptors to significantly extend the applicability machine learning (ML) in material screening and mapping energy landscape for multicomponent systems. These allow differentiating between structural prototypes, which is not possible using commonly used chemical-only descriptors. Specifically, we demonstrate that combination pairwise radial, nearest-neighbor, bond-angle, dihedral-angle, core-charge distributions plays an important role...
The Open Databases Integration for Materials Design (OPTIMADE) consortium has designed a universal application programming interface (API) to make materials databases accessible and interoperable. We outline the first stable release of specification, v1.0, which is already supported by many leading several software packages. illustrate advantages OPTIMADE API through worked examples on each public that support full specification.
Abstract We develop a multi-step workflow for the discovery of conventional superconductors, starting with Bardeen–Cooper–Schrieffer inspired pre-screening 1736 materials high Debye temperature and electronic density states. Next, we perform electron-phonon coupling calculations 1058 them to establish large systematic database BCS superconducting properties. Using McMillan-Allen-Dynes formula, identify 105 dynamically stable transition temperatures, T C ≥ 5 K. Additionally, analyze trends in...
High-throughput density functional theory (DFT) calculations allow for a systematic search conventional superconductors. With the recent interest in two-dimensional (2D) superconductors, we used high-throughput workflow to screen over 1000 2D materials JARVIS-DFT database and performed electron-phonon coupling calculations, using McMillan-Allen-Dynes formula calculate superconducting transition temperature (Tc) 165 of them. Of these materials, identify 34 dynamically stable structures with...
In this work, we present the ChemNLP library that can be used for (1) curating open access datasets materials and chemistry literature, developing comparing traditional machine learning, transformers graph neural network models (2) classifying clustering texts, (3) named entity recognition large-scale text-mining, (4) abstractive summarization generating titles of articles from abstracts, (5) text generation suggesting abstracts titles, (6) integration with density functional theory dataset...
We develop a computational database, website applications (web-apps), and machine-learning (ML) models to accelerate the design discovery of two-dimensional (2D) heterostructures. Using density functional theory (DFT) based lattice parameters electronic band energies for 674 nonmetallic exfoliable 2D materials, we generate 226 779 possible bilayer classify these heterostructures into type-I, -II, -III systems according Anderson's rule, which is on relative alignments noninteracting...
Finding new superconductors with a high critical temperature (Tc) has been challenging task due to computational and experimental costs. We present diffusion model inspired by the computer vision community generate unique structures chemical compositions. Specifically, we used crystal variational autoencoder (CDVAE) along atomistic line graph neural network (ALIGNN) pretrained models Joint Automated Repository for Various Integrated Simulations (JARVIS) superconducting database of density...