Predicting materials properties with generative models: applying generative adversarial networks for heat flux generation

Qi Kong,Yasushi Shibuta

DOI: https://doi.org/10.1088/1361-648x/ad258b

2024-02-04

Journal of Physics Condensed Matter

Abstract:In the realm of materials science, the integration of machine learning techniques has ushered in a transformative era. This study delves into the innovative application of generative adversarial networks (GANs) for generating heat flux data, a pivotal step in predicting lattice thermal conductivity within metallic materials. Leveraging GANs, this research explores the generation of meaningful heat flux data, which has a high degree of similarity with that calculated by molecular dynamics (MD) simulations. This study demonstrates the potential of artificial intelligence (AI) in understanding the complex physical meaning of data in materials science. By harnessing the power of such AI to generate data that is previously attainable only through experiments or simulations, new opportunities arise for exploring and predicting properties of materials.

physics, condensed matter

What problem does this paper attempt to address?

The main objective of this paper is to explore how to use Generative Adversarial Networks (GANs) to generate heat flux data similar to that obtained from Molecular Dynamics (MD) simulations, and further use this generated data to predict the lattice thermal conductivity of metallic materials. Specifically, the researchers employed GANs to generate heat flux data, which is crucial in materials science for understanding the thermal conduction properties of materials. Heat flux data can be obtained through MD simulations, and the lattice thermal conductivity can be calculated using the Green-Kubo formula. However, MD simulations can only capture the part of thermal conduction caused by lattice vibrations and cannot directly account for the contribution of electrons to thermal conductivity. Therefore, the thermal conductivity obtained from MD simulations is usually lower than the actual measured value. To overcome this limitation, the research team attempted to use GANs to generate heat flux data that closely resembles the results of MD simulations. They demonstrated that GANs could generate heat flux data with similar amplitude and oscillation patterns, and the consistency in frequency characteristics between the generated data and MD simulation data was confirmed through Fast Fourier Transform (FFT) analysis. Finally, the lattice thermal conductivity calculated using the generated heat flux data based on the Green-Kubo formula was very close to the results of MD simulations, indicating that the data generated by GANs can be effectively used to predict the thermal conductivity of materials. In summary, this paper aims to address the problem of how to use GANs to generate high-quality heat flux data to more accurately predict the lattice thermal conductivity of metallic materials.

Predicting materials properties with generative models: applying generative adversarial networks for heat flux generation

Predicting the Effective Thermal Conductivities of Composite Materials and Porous Media by Machine Learning Methods

FluxGAN: A Physics-Aware Generative Adversarial Network Model for Generating Microstructures That Maintain Target Heat Flux

Predicting thermal stress in binary composites through advanced generative adversarial networks

Exploring lattice thermal conductivity models via interpretable deep learning to accelerate the discovery of novel materials

Learning Representations of Inorganic Materials from Generative Adversarial Networks

Applications of Generative Adversarial Networks in Materials Science

Quick Prediction of Complex Temperature Fields Using Conditional Generative Adversarial Networks

Revolutionising inverse design of magnesium alloys through generative adversarial networks

LatticeML: A data-driven application for predicting the effective Young Modulus of high temperature graph based architected materials

Mechanistic data-driven prediction of as-built mechanical properties in metal additive manufacturing

Deep learning for synthetic microstructure generation in a materials-by-design framework for heterogeneous energetic materials

Generative artificial intelligence and its applications in materials science: Current situation and future perspectives

Predicting lattice thermal conductivity from fundamental material properties using machine learning techniques

Generative adversarial network–based real-time temperature prediction model for heating stage of electric arc furnace

ScGAN: a generative adversarial network to predict hypothetical superconductors

Predictive Modeling of High-Entropy Alloys and Amorphous Metallic Alloys Using Machine Learning

Generative deep learning as a tool for inverse design of high-entropy refractory alloys

Improving direct physical properties prediction of heterogeneous materials from imaging data via convolutional neural network and a morphology-aware generative model

Machine Learning Models for the Lattice Thermal Conductivity Prediction of Inorganic Materials

Functional Output Regression for Machine Learning in Materials Science