Establishing a data-driven pipeline for the discovery of novel materials requires the engineering of material features that can be feasibly calculated and can be applied to predict a material's target properties. Here we propose a new class of descriptors for describing crystal structures, which we term Robust One-Shot Ab initio (ROSA) descriptors. ROSA is computationally cheap and is shown to accurately predict a range of material properties. These simple and intuitive class of descriptors are generated from the energetics of a material at a low level of theory using an incomplete ab initio calculation. We demonstrate how the incorporation of ROSA descriptors in ML-based property prediction leads to accurate predictions over a wide range of crystals, amorphized crystals, metal-organic frameworks and molecules. We believe that the low computational cost and ease of use of these descriptors will significantly improve ML-based predictions.
Compositional disorder is common in crystal compounds. In these compounds, some atoms are randomly distributed at some crystallographic sites. For such compounds, randomness forms many non-identical independent structures. Thus, calculating the energy of all structures using ordinary quantum ab initio methods can be significantly time-consuming. Machine learning can be a reliable alternative to ab initio methods. We calculate the energy of these compounds with an accuracy close to that of density functional theory calculations in a considerably shorter time using machine learning. In this stud…
Active-Learning Inspired $\textit{Ab Initio}$ Theory-Experiment Loop Approach for Management of Material Defects: Application to Superconducting Qubits
In conventional chemisorption model, the d-band center theory (augmented sometimes with the upper edge of d-band for imporved accuarcy) plays a central role in predicting adsorption energies and catalytic activity as a function of d-band center of the solid surfaces, but it requires density functional calculations that can be quite costly for large scale screening purposes of materials. In this work, we propose to use the d-band width of the muffin-tin orbital theory (to account for local coordination environment) plus electronegativity (to account for adsorbate renormalization) as a simple se…
A main goal of data-driven materials research is to find optimal low-dimensional descriptors, allowing us to predict a physical property, and to interpret them in a human-understandable way. In this work, we advance methods to identify descriptors out of a large pool of candidate features by means of compressed sensing. To this extent, we develop schemes for engineering appropriate candidate features that are based on simple basic properties of building blocks that constitute the materials and that are able to represent a multi-component system by scalar numbers. Cross-validation based feature…
Persistent homology-based descriptor for machine-learning potential of amorphous structures
The Institute of Scientific and Industrial Research, Osaka University（大阪大学产业科学研究所）；Institute for Molecular Science（分子科学研究所）；JST, PRESTO（日本科学技术振兴机构PRESTO项目）；Center for Artificial Intelligence and Mathematical Data Science, Okayama University（冈山大学人工智能与数学数据科学中心）；The University of Tokyo（东京大学）
Lawrence Livermore National Laboratory（劳伦斯利弗莫尔国家实验室）
One of the grand challenges of utilizing machine learning for the discovery of innovative new polymers lies in the difficulty of accurately representing the complex structures of polymeric materials. Although a wide array of hand-designed polymer representations have been explored, there has yet to be an ideal solution for how to capture the periodicity of polymer structures, and how to develop polymer descriptors without the need for human feature design. In this work, we tackle these problems through the development of our periodic polymer graph representation. Our pipeline for polymer prope…
Historically, materials discovery has been driven by a laborious trial-and-error process. The growth of materials databases and emerging informatics approaches finally offer the opportunity to transform this practice into data- and knowledge-driven rational design. By using data from the AFLOW repository for high-throughput ab-initio calculations, we have generated Quantitative Materials Structure-Property Relationship (QMSPR) models to predict eight critical electronic and thermomechanical materials properties, such as the metal/insulator classification, band gap energy, bulk and shear moduli…
The lack of reliable methods for identifying descriptors - the sets of parameters capturing the underlying mechanisms of a materials property - is one of the key factors hindering efficient materials development. Here, we propose a systematic approach for discovering descriptors for materials properties, within the framework of compressed-sensing based dimensionality reduction. SISSO (sure independence screening and sparsifying operator) tackles immense and correlated features spaces, and converges to the optimal solution from a combination of features relevant to the materials' property of in…
We evaluate the performance of four machine learning methods for modeling and predicting FCC solute diffusion barriers. More than 200 FCC solute diffusion barriers from previous density functional theory (DFT) calculations served as our dataset to train four machine learning methods: linear regression (LR), decision tree (DT), Gaussian kernel ridge regression (GKRR), and artificial neural network (ANN). We separately optimize key physical descriptors favored by each method to model diffusion barriers. We also assess the ability of each method to extrapolate when faced with new hosts with limit…
Lattice thermal conductivities (LTC) for a subset of polymer crystals from the Polymer Genome Library were investigated to explore high LTC polymer systems. We employed a first-principles approach to evaluating phonon lifetimes within the third-order perturbation theory combined with density functional theory, and then solved the linearized Boltzmann transport equation with single-mode relaxation time approximated by the computed lifetime. Typical high LTC polymer systems, polyethylene (PE) crystal and fiber, were benchmarked, which is reasonably consistent with previous references, validating…
The (Cd, Be)Te semiconductor alloy that exhibits a dramatic mismatch in bond covalency and stiffness clarifying its vibrational and mechanical properties is used as a benchmark to test the limits of the percolation model (PM) worked out to explain the complex Raman spectra of the related but less contrasted (Zn, Be) chalcogenides. The test is done by way of experiment (x smaller than 0.11) combining Raman scattering with X ray diffraction at high pressure and ab initio calculations (x around 0, 0.5 and 1). The (macroscopic) bulk modulus B drops below the CdTe value on minor Be incorporation, a…
Direction-aware topological descriptors for elastic stiffness tensor prediction in porous materials
Dioscuri Center in Topological Data Analysis, Institute of Mathematics, Polish Academy of Sciences（波兰科学院数学研究所拓扑数据分析迪奥苏里中心）；Institute of Experimental Physics, University of Wrocław（弗罗茨瓦夫大学实验物理研究所）；Faculty of Pure and Applied Mathematics, Wrocław University of Science and Technology（弗罗茨瓦夫科技大学纯与应用数学学院）；Faculty of Mathematics and Computer Science, Adam Mickiewicz University（亚当·密茨凯维奇大学数学与计算机科学学院）；IMDEA Materials Institute（IMDEA材料研究所）
Ab initio prediction of the variation of properties in the configurational space of solid solutions is computationally very demanding. We present an approach to accelerate these predictions via a combination of density functional theory and machine learning, using the cubic spinel nitride GeSn$_2$N$_4$ as a case study, exploring how formation energy and electronic bandgap are affected by configurational variations. Furthermore, we demonstrate the utility of applying explainable machine learning to understand the crystal chemistry origins of the trends that we observe. Different configuration d…
We have developed a method that can analyze large random grain boundary (GB) models with the accuracy of density functional theory (DFT) calculations using active learning. It is assumed that the atomic energy is represented by the linear regression of the atomic structural descriptor. The atomic energy is obtained through DFT calculations using a small cell extracted from a huge GB model, called replica DFT atomic energy. The uncertainty reduction (UR) approach in active learning is used to efficiently collect the training data for the atomic energy. In this approach, atomic energy is not req…
Materials Expert-Artificial Intelligence for Materials Discovery
Department of Physics, Cornell University, Ithaca, NY 14853, USA；Department of Chemistry, Princeton University, Princeton, New Jersey 08544, United States；Department of Chemistry, North Carolina State University, Raleigh, NC, USA；New York University, New York, NY 10011, USA；Fraunhofer Research Institution for Materials Recycling；Department of Physics, Ewha Womans University, Seoul, South Korea
Computing vibrational free energies ($F_{vib}$) and entropies ($S_{vib}$) has posed a long standing challenge to the high-throughput ab initio investigation of finite temperature properties of solids. Here we use machine-learning techniques to efficiently predict $F_{vib}$ and $S_{vib}$ of crystalline compounds in the Inorganic Crystal Structure Database. By employing descriptors based simply on the chemical formula and using a training set of only 300 compounds, mean absolute errors of less than 0.04 meV/K/atom (15 meV/atom) are achieved for $S_{vib}$ ($F_{vib}$), whose values are distributed…
We propose a scheme to construct predictive models for Hamiltonian matrices in atomic orbital representation from ab initio data as a function of atomic and bond environments. The scheme goes beyond conventional tight binding descriptions as it represents the ab initio model to full order, rather than in two-centre or three-centre approximations. We achieve this by introducing an extension to the Atomic Cluster Expansion (ACE) descriptor that represents Hamiltonian matrix blocks that transform equivariantly with respect to the full rotation group. The approach produces analytical linear models…
We combined descriptor-based analytical models for stiffness-matrix and elastic-moduli with mean-field methods to accelerate assessment of technologically useful properties of high-entropy alloys, such as strength and ductility. Model training for elastic properties uses Sure-Independence Screening (SIS) and Sparsifying Operator (SO) method yielding an optimal analytical model, constructed with meaningful atomic features to predict target properties. Computationally inexpensive analytical descriptors were trained using a database of the elastic properties determined from density functional the…
We study the structural, electronic, and magnetic properties of Co$_2$ZAl compounds employing a pseudopotential electronic bandstructure method. The stability of the compounds is established through the formation and cohesive energy calculations. The effect of the lattice parameter variation on the electronic and magnetic properties of the compounds is investigated and meticulous explanation is provided for the observed behavior. The variation of the individual spin magnetic moments and the stability of the total spin magnetic moment during the expansion and contraction of the lattice paramete…
Despite the fast development of computational materials modelling, theoretical description of macroscopic elastic properties of textured polycrystalline aggregates starting from basic principles remains a challenging task. In this communication we use a supercell-based approach to obtain the elastic properties of random solid solution cubic ZrAlN system as a function of the metallic sublattice composition and texture descriptors. The employed special quasi-random structures are optimised not only with respect to short range order parameters, but also to make the three cubic directions $[1\,0\,…
Combining feature-based approaches with graph neural networks and symbolic regression for synergistic performance and interpretability
Laboratory of Applied Materials and Interfaces, Federal University of Rio Grande do Sul（南里奥格兰德联邦大学应用材料与界面实验室）；Institute of Condensed Matter and Nanosciences, Université Catholique de Louvain（鲁汶天主教大学凝聚态与纳米科学研究所）；Ruhr-Universität Bochum（波鸿鲁尔大学）
Traditional materials discovery approaches - relying primarily on laborious experiments - have controlled the pace of technology. Instead, computational approaches offer an accelerated path: high-throughput exploration and characterization of virtual structures. These ventures, performed by automated ab-initio frameworks, have rapidly expanded the volume of programmatically-accessible data, cultivating opportunities for data-driven approaches. Herein, a collection of robust characterization methods are presented, implemented within the Automatic Flow Framework for Materials Discovery (AFLOW), …
The polaron is the archetypal example of a quasiparticle emerging from the interaction between fermionic and bosonic fields in quantum field theory. In crystalline solids, polarons are formed when electrons and holes become dressed by the quanta of lattice vibrations. While experimental signatures of polarons in bulk three-dimensional materials abound, only rarely have polarons been observed in two-dimensional atomic crystals. Here, we shed light on this asymmetry by developing a quantitative ab initio theory of polarons in atomically-thin crystals. Using this conceptual framework, we unravel …
In this paper we develop the stability rules for NASICON structured materials, as an example of compounds with complex bond topology and composition. By applying machine learning to the ab-initio computed phase stability of 3881 potential NASICONs we can extract a simple two-dimensional descriptor that is extremely good at separating stable from unstable NASICONS. This machine-learned "tolerance factor" contains information on the Na content, the radii and electronegativities of the elements, and the Madelung energy. We test the predictive capability of this approach by selecting six predicted…
