High-throughput discovery of novel cubic crystal materials using deep generative neural networks

Zhao, Yong; Al-Fahdi, Mohammed; Hu, Ming; Siriwardane, Edirisuriya MD; Song, Yuqi; Nasiri, Alireza; Hu, Jianjun

doi:10.48550/arxiv.2102.01880

Cited by 4 publications

(11 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work, we use our recently developed CubicGAN algorithm [46] to generate 10 million hypothetical ternary cubic crystal structures of three space groups (221,225, 216) which are reduced to 2.5 million unique candidate cubic structures. With such a high volume of candidates, how to find the stable and synthesizable ones is almost like finding the needle in the haystack.…”

Section: The Framework For Generative Design Of Materialsmentioning

confidence: 99%

See 1 more Smart Citation

Semi-supervised teacher-student deep neural network for materials discovery

Daniel¹,

Siriwardane²,

Zhao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Data driven generative machine learning models have recently emerged as one of the most promising approaches for new materials discovery. While the generator models can generate millions of candidates, it is critical to train fast and accurate machine learning models to filter out stable, synthesizable materials with desired properties. However, such efforts to build supervised regression or classification screening models have been severely hindered by the lack of unstable or unsynthesizable samples, which usually are not collected and deposited in materials databases such as ICSD and Materials Project (MP). At the same time, there are a significant amount of unlabelled data available in these databases. Here we propose a semi-supervised deep neural network (TSDNN) model for high-performance formation energy and synthesizability prediction, which is achieved via its unique teacher-student dual network architecture and its effective exploitation of the large amount of unlabeled data. For formation energy based stability screening, our semi-supervised classifier achieves an absolute 10.3% accuracy improvement compared to the baseline CGCNN regression model. For synthesizability prediction, our model significantly increases the baseline PU learning's true positive rate from 87.9% to 97.9% using 1/49 model parameters. To further prove the effectiveness of our models, we combined our TSDNN-energy and TSDNN-synthesizability models with our CubicGAN generator to discover novel stable cubic structures. Out of 1000 recommended candidate samples by our models, 512 of them have negative formation energies as validated by our DFT formation energy calculations. Our experimental results show that our semi-supervised deep neural networks can significantly improve the screening accuracy in large-scale generative materials design.

show abstract

Section: The Framework For Generative Design Of Materialsmentioning

confidence: 99%

“…CubicGAN [46] is a generative adversarial network based model for generating novel cubic crystal structures. Cubic-GAN reports that when generating 10 million virtual cubic crystal structures, most of materials in training datasets, Materials Project and ICSD can rediscovered.…”

Section: Generation Of Candidate Cubic Structures For Screeningmentioning

confidence: 99%

Semi-supervised teacher-student deep neural network for materials discovery

Daniel¹,

Siriwardane²,

Zhao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Another promising approach to design solid materials beyond known crystal structure prototypes is generative deep learning models [36], which can learn data distribution (knowledge of forming stable crystal structures) from known materials and then sample from it to generate novel materials. Variational Auto-encoder (VAE) [19] and Generative Adversarial Network (GAN) [12] are two popular generative models used to generate materials.…”

Section: Introductionmentioning

confidence: 99%

“…However, it remains uncertain whether CrystalGAN can be extended to produce more complex crystals. Both GANCSP [18] and CubicGAN [36] use a "point cloud" (containing fractional coordinates, element properties, and lattice parameters) as inputs to build a model that generates crystals conditioned on composition or both composition and space group. The difference between them is GANCSP can only generate structures of the Mg-Mn-O system but Cu-bicGAN can generate more diverse systems under three space groups.…”

Section: Introductionmentioning

confidence: 99%

Physics Guided Deep Learning for Generative Design of Crystal Materials with Symmetry Constraints

Zhao¹,

Siriwardane²,

Wu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Discovering new materials is a long-standing challenging task that is critical to the progress of human society. Conventional approaches such as trial-and-error experiments and computational simulations are labor-intensive or costly with their success heavily depending on experts' heuristics. Recently deep generative models have been successfully proposed for materials generation by learning implicit knowledge from known materials datasets, with performance however limited by their confinement to a special material family or failing to incorporate physical rules into the model training process. Here we propose a Physics Guided Crystal Generative Model (PGCGM) for new materials generation, which captures and exploits the pairwise atomic distance constraints among neighbor atoms and symmetric geometric constraints. By augmenting the base atom sites of materials, our model can generates new materials of 20 space groups. With atom clustering and merging on generated crystal structures, our method increases the generator's validity by 8 times compared to one of the baselines and by 143% compared to the previous CubicGAN along with its superiority in properties distribution and diversity. We further validated our generated candidates by Density Functional Theory (DFT) calculation, which successfully optimized/relaxed 1869 materials out of 2000, of which 39.6% are with negative formation energy, indicating their stability.

show abstract

“…Compared to the vast chemical space of crystal materials, the known crystal structures ( 200,000) as deposited in the ICSD and Materials Project database are quite limited. Recently, we proposed a generative machine learning model CubicGAN [22] for automated generation of cubic crystal structures, allowing us to discover hundreds of new prototype cubic materials. However, that approach is currently only limited to generate cubic structures with special coordinates.…”

Section: Introductionmentioning

confidence: 99%

Crystal structure prediction using age-fitness multi-objective genetic algorithm and coordination number constraints

Yang,

Siriwardane,

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Crystal structure prediction (CSP) has emerged as one of the most important approaches for discovering new materials. CSP algorithms based on evolutionary algorithms and particle swarm optimization have discovered a great number of new materials. However, these algorithms based on ab initio calculation of free energy are inefficient. Moreover, they have severe limitations in terms of scalability. We recently proposed a promising crystal structure prediction method based on atomic contact maps, using global optimization algorithms to search for the Wyckoff positions by maximizing the match between the contact map of the predicted structure and the contact map of the true crystal structure. However, our previous contact map based CSP algorithms have two major limitations: (1) the loss of search capability due to getting trapped in local optima; (2) it only uses the connection of atoms in the unit cell to predict the crystal structure, ignoring the chemical environment outside the unit cell, which may lead to unreasonable coordination environments. Herein we propose a novel multi-objective genetic algorithms for contact map-based crystal structure prediction by optimizing three objectives, including contact map match accuracy, the individual age, and the coordination number match. Furthermore, we assign the age values to all the individuals of the GA and try to minimize the age aiming to avoid the premature convergence problem. Our experimental results show that compared to our previous CMCrystal algorithm, our multi-objective crystal structure prediction algorithm (CMCrystalMOO) can reconstruct the crystal structure with higher quality and alleviate the problem of premature convergence.

show abstract

High-throughput discovery of novel cubic crystal materials using deep generative neural networks

Cited by 4 publications

References 35 publications

Semi-supervised teacher-student deep neural network for materials discovery

Semi-supervised teacher-student deep neural network for materials discovery

Physics Guided Deep Learning for Generative Design of Crystal Materials with Symmetry Constraints

Crystal structure prediction using age-fitness multi-objective genetic algorithm and coordination number constraints

Contact Info

Product

Resources

About