Crystal structure prediction in a continuous representative space

Chang, Kee-Joo

doi:10.1016/j.commatsci.2021.110436

Cited by 17 publications

(15 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While the peer protein structure prediction problem has recently been almost solved by the deep-learning-based AlphaFold and RossettaFold algorithms, the CSP problem remains elusive for a majority of categories of compositions. There are mainly three types of CSP approaches including the ab initio based global optimization − as reviewed in ref , machine-learning-based prediction, , and template-based elemental substitution . The first approach instead depends on computationally expensive density functional theory (DFT) calculations and is applicable to only small chemical systems.…”

Section: Introductionmentioning

confidence: 99%

TCSP: a Template-Based Crystal Structure Prediction Algorithm for Materials Discovery

Wei

Siriwardane

et al. 2022

Inorg. Chem.

View full text Add to dashboard Cite

Fast and accurate crystal structure prediction (CSP) algorithms and web servers are highly desirable for the exploration and discovery of new materials out of the infinite chemical design space. However, currently, the computationally expensive first-principles calculation-based CSP algorithms are applicable to relatively small systems and are out of reach of most materials researchers. Several teams have used an element substitution approach for generating or predicting new structures, but usually in an ad hoc way. Here we develop a template-based crystal structure prediction (TCSP) algorithm and its companion web server, which makes this tool accessible to all materials researchers. Our algorithm uses elemental/chemical similarity and oxidation states to guide the selection of template structures and then rank them based on the substitution compatibility and can return multiple predictions with ranking scores in a few minutes. A benchmark study on the 98290 formulas of the Materials Project database using leave-one-out evaluation shows that our algorithm can achieve high accuracy (for 13145 target structures, TCSP predicted their structures with root-mean-square deviation < 0.1) for a large portion of the formulas. We have also used TCSP to discover new materials of the Ga–B–N system, showing its potential for high-throughput materials discovery. Our user-friendly web app TCSP can be accessed freely at on our MaterialsAtlas.org web app platform.

show abstract

Section: Introductionmentioning

confidence: 99%

TCSP: a Template-Based Crystal Structure Prediction Algorithm for Materials Discovery

Wei

Siriwardane

et al. 2022

Inorg. Chem.

View full text Add to dashboard Cite

show abstract

“…Deep generative frameworks employing neural networks (NNs) have proven to be an invaluable tool in this context . Notably, these comprise a large zoo of approaches, including variational autoencoders (VAEs), − ,,,− generative adversarial networks (GANs), ,,− ,, and reinforcement learning (RL). , Given this wide range of approaches, it is a priori difficult to decide which method should be used for a new inverse design task.…”

Section: Introductionmentioning

confidence: 99%

Assessing Deep Generative Models in Chemical Composition Space

et al. 2022

View full text Add to dashboard Cite

The computational discovery of novel materials has been one of the main motivations behind research in theoretical chemistry for several decades. Despite much effort, this is far from a solved problem, however. Among other reasons, this is due to the enormous space of possible structures and compositions that could potentially be of interest. In the case of inorganic materials, this is exacerbated by the combinatorics of the periodic table since even a single-crystal structure can in principle display millions of compositions. Consequently, there is a need for tools that enable a more guided exploration of the materials design space. Here, generative machine learning models have recently emerged as a promising technology. In this work, we assess the performance of a range of deep generative models based on reinforcement learning, variational autoencoders, and generative adversarial networks for the prototypical case of designing Elpasolite compositions with low formation energies. By relying on the fully enumerated space of 2 million main-group Elpasolites, the precision, coverage, and diversity of the generated materials are rigorously assessed. Additionally, a hyperparameter selection scheme for generative models in chemical composition space is developed.

show abstract

“…34 Moreover, a much larger variety of chemical compositions are possible for inorganic systems. In the prevailing data scarcity, inverse design of inorganic materials has therefore usually focused more on structural prediction in limited composition spaces, [35][36][37][38][39][40] or compositional optimization with fixed structural prototypes. [41][42][43] However, even if trained on a limited subset of structure types, it has been shown that these models are able to generalize to new structure types that were not included in the training process.…”

Section: Introductionmentioning

confidence: 99%

“…34 In the above examples, deep generative frameworks employing neural networks (NNs) have proven to be an invaluable tool. 44 Notably, these comprise a large zoo of approaches, including variational autoencoders (VAEs), [34][35][36]39,42,[45][46][47] generative adversarial networks (GANs) 37,38,[41][42][43]48,49 and reinforcement learning (RL). 50,51 Given this wide range of approaches it is a priori difficult to decide which method should be used for a new inverse design task.…”

Section: Introductionmentioning

confidence: 99%

Assessing Deep Generative Models in Chemical Composition Space

Türk

Landini

Künkel

et al. 2022

Preprint

View full text Add to dashboard Cite

The computational discovery of novel materials has been one of the main motivations behind research in theoretical chemistry for several decades. Despite much effort, this is far from a solved problem, however. Among other reasons, this is due to the enormous space of possible structures and compositions that could potentially be of interest. In the case of inorganic materials, this is exacerbated by the combinatorics of the periodic table, since even a single crystal structure can in principle display millions of compositions. Consequently, there is a need for tools that enable a more guided exploration of the materials design space. Here, generative machine learning (ML) models have recently emerged as a promising technology. In this work, we assess the performance of a range of deep generative models based on Reinforcement Learning (RL), Variational Autoencoders (VAE) and Generative Adversarial Networks (GAN) for the prototypical case of designing Elpasolite compositions with low formation energies. By relying on the fully enumerated space of 2~million main group Elpasolites, the precision, coverage and diversity of the generated materials is rigorously assessed. Additionally, a hyperparameter selection scheme for generative models in chemical composition space is developed.

show abstract

Crystal structure prediction in a continuous representative space

Cited by 17 publications

References 49 publications

TCSP: a Template-Based Crystal Structure Prediction Algorithm for Materials Discovery

TCSP: a Template-Based Crystal Structure Prediction Algorithm for Materials Discovery

Assessing Deep Generative Models in Chemical Composition Space

Assessing Deep Generative Models in Chemical Composition Space

Contact Info

Product

Resources

About