Materials Properties Prediction (MAPP): Empowering the prediction of material properties solely based on chemical formulas
9 November 2023
MAPP framework predicts material properties from only chemical formulas
Uses graph neural networks to represent materials for property prediction
Models trained on large datasets can rapidly predict properties
Public web application allows open access to material property predictions
Framework aims to accelerate materials discovery and design
Predicting material properties from chemical formulas
This paper introduces a machine learning framework called MAPP that can predict key properties of materials based only on their chemical formulas. It uses graph neural networks to represent materials as graphs with elements as nodes. Training on large datasets, the models can rapidly predict properties like melting point and bulk modulus for any material formula. The framework is publicly accessible, allowing material scientists to leverage AI for accelerated discovery.
No comments yet, be the first to start the conversation...
Sign up to comment on this paper