MAterials Properties Prediction (MAPP)


built on experimental data
empowered by artificial intelligence
enhanced by ab initio calculations

PropertiesDataset SizeR2 Score (training/testing)RMSE (training/testing)Status
Consortium modelDeployed
Melting Temperature9,1980.995/0.9850/100 KDeployed
Bulk Modulus4,2360.997/0.974/14 GPaDeployed
Volume49,2130.985/0.9750.942/1.177 Ang^3/atomDeployed
Superconductor Critical Temperature12,4480.92/0.908.5/9.0 KDeployed
Crystal Structure7,882AUC: 0.998/0.961F1 Score: 0.86/0.66Deployed
Heat Capacity2,0740.72/0.630.66/0.79 J mol-1K-1In Progress
Heat of Fusion1,2330.94/0.890.4/0.6 kJ mol-1In Progress
Thermal Expansion   Planned
Calphad model excess parameter LA,B;0790 0.96/0.85 18/35 kJ mol-1Deployed

Not everyone is an expert on machine learning
We share these models with you, so…
MAterials Properties Prediction is just a few clicks away


Consortium, an ensemble of 30 GNN models, accepts up to 4 elements

Try La_2Zr_2O_7, or La_2O_3(ZrO_2)_2, or ZrO_2, or HfC_0.93, or Ni_10Fe_72Cr_18

Cite this model:

Si-Da Xue and Qi-Jun Hong, Materials Properties Prediction (MAPP): Empowering the prediction of material properties solely based on chemical formulas, arXiv, 2023. Download.

Qi-Jun Hong, A melting temperature database and a neural network model for melting temperature prediction, arXiv, 2021. Download.

Qi-Jun Hong, Sergey V Ushakov, Axel van de Walle, and Alexandra Navrotsky, Melting temperature prediction using a graph neural network model: From ancient minerals to new materials, PNAS, 2022. Download.

This model is currently deployed on Microsoft Azure and the Research Computing facilities at ASU. Due to network limit, computation may take up to 10 seconds.

About this model

This model is based on Graph Neural Network (GNN) and Residual Neural Network (ResNet).
This ensemble model is based on bootstrap aggregating (bagging).

Video Introduction