Neural network force fields for simple metals and semiconductors: construction and application to the calculation of phonons and melting temperatures

Marques, Mário R. G.; Wolff, Jakob; Steigemann, Conrad; Marques, Miguel A. L.

doi:10.1039/c8cp05771k

Cited by 30 publications

(17 citation statements)

References 95 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the computational cost inherent to these classical approximations have limited the size, flexibility, and extensibility of the studies. Larger searches on relevant chemical patterns, have been successfully conducted since several research groups have developed ML models and algorithms to predict chemical properties using training data generated by DFT, which have also contributed to the increase of public collections of molecules coupled with vibrational, thermodynamic and DFT computed electronic properties (e.g., Behler and Parrinello, 2007;Rupp et al, 2012;Behler, 2016;Hegde and Bowen, 2017;Pronobis et al, 2018;Chandrasekaran et al, 2019;Iype and Urolagin, 2019;Marques et al, 2019;Schleder et al, 2019).…”

Section: Co-occurring Machine-learning Contributions In Chemical Sciementioning

confidence: 99%

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

2019

View full text Add to dashboard Cite

Computational Chemistry is currently a synergistic assembly between ab initio calculations, simulation, machine learning (ML) and optimization strategies for describing, solving and predicting chemical data and related phenomena. These include accelerated literature searches, analysis and prediction of physical and quantum chemical properties, transition states, chemical structures, chemical reactions, and also new catalysts and drug candidates. The generalization of scalability to larger chemical problems, rather than specialization, is now the main principle for transforming chemical tasks in multiple fronts, for which systematic and cost-effective solutions have benefited from ML approaches, including those based on deep learning (e.g. quantum chemistry, molecular screening, synthetic route design, catalysis, drug discovery). The latter class of ML algorithms is capable of combining raw input into layers of intermediate features, enabling bench-to-bytes designs with the potential to transform several chemical domains. In this review, the most exciting developments concerning the use of ML in a range of different chemical scenarios are described. A range of different chemical problems and respective rationalization, that have hitherto been inaccessible due to the lack of suitable analysis tools, is thus detailed, evidencing the breadth of potential applications of these emerging multidimensional approaches. Focus is given to the models, algorithms and methods proposed to facilitate research on compound design and synthesis, materials design, prediction of binding, molecular activity, and soft matter behavior. The information produced by pairing Chemistry and ML, through data-driven analyses, neural network predictions and monitoring of chemical systems, allows (i) prompting the ability to understand the complexity of chemical data, (ii) streamlining and designing experiments, (ii) discovering new molecular targets and materials, and also (iv) planning or rethinking forthcoming chemical challenges. In fact, optimization engulfs all these tasks directly.

show abstract

Section: Co-occurring Machine-learning Contributions In Chemical Sciementioning

confidence: 99%

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

2019

View full text Add to dashboard Cite

show abstract

“…25 3.1 Feature selection and data normalization ANN can discover relationships between inputs and outputs of a system without understanding sophisticated function mechanism. 26 Here, we apply ANN for analyzing the relationship between the structure of carbon materials and performance of supercapacitors. There are many kinds of structure features affecting the electrochemical performance of supercapacitor.…”

Section: Ann Modelmentioning

confidence: 99%

Modifying the microstructure of algae-based active carbon and modelling supercapacitors using artificial neural networks

Wang

Yan

et al. 2019

RSC Adv.

View full text Add to dashboard Cite

show abstract

“…Over the last decade, impressive progress was made in designing potentials from electronic structure calculations using supervised Machine Learning (ML) methods [27][28][29][30][31][32][33]. There are now standard libraries for the ML training [34] that can be used in combination with molecular dynamics (MD) simulation packages [35] such as LAMMPS [36] or in combination with workflow softwares such as ASE [37].…”

Section: Introductionmentioning

confidence: 99%

“…ML potentials were initially trained using the DFT energies starting from the work of Behler and Parrinello on bulk Si [40]. It was subsequently pointed out that the learning process could benefit from a wealth of additional information if the three components of the force and six components of the stress per atom are taken into account [27,33], while one has only a single energy value per simulated configuration. Still in some works, only the forces have been used for the training showing that properties like the vibrational properties in the solid states can be reproduced, but they remain insufficient to get full account of thermodynamic quantities [44,50].…”

Section: Introductionmentioning

confidence: 99%

Interatomic machine learning potentials for aluminium: application to solidification phenomena

Jakse¹,

Sandberg²,

Saliou³

et al. 2022

Preprint

View full text Add to dashboard Cite

In studying solidification process by simulations on the atomic scale, modeling crystal nucleation or amorphisation require the construction of interatomic interactions that are able to reproduce the properties of both the solid and the liquid states. Taking into account rare nucleation events or structural relaxation under deep undercooling conditions requires much larger length scales and longer time scales than those achievable by ab initio molecular dynamics (AIMD). This problem is addressed by means of classical MD simulations using a high dimensional neural network potential trained on a relevant set of configurations generated by AIMD and applied to elemental aluminium.The dataset contains various crystalline structures and liquid states at different pressures, including their time fluctuations in a wide range of temperatures. The resulting potential is shown to be efficient to reproduce the basic structural, dynamics and thermodynamic quantities in the liquid and undercooled states with an accuracy close to the ab initio one, without the need to include neither explicitly the forces nor all kind of configurations in the training procedure. The early stages of the crystallization is further investigated on a scale much larger than the ab initio allowing us to unravel features of the homogeneous nucleation mechanisms.

show abstract

Neural network force fields for simple metals and semiconductors: construction and application to the calculation of phonons and melting temperatures

Abstract: We present a practical procedure to obtain reliable and unbiased neural network based force fields for solids.

Cited by 30 publications

References 95 publications

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

Modifying the microstructure of algae-based active carbon and modelling supercapacitors using artificial neural networks

Interatomic machine learning potentials for aluminium: application to solidification phenomena

Contact Info

Product

Resources

About