About
- Royce supports world-recognised excellence in UK materials research, accelerating commercial exploitation of innovations and delivering positive economic and societal impact for the UK.
  
  More About Royce
- About Royce
Research
- Research undertaken at Royce tackles some of the most pressing challenges facing today’s society, from providing energy for future cities to decarbonisation and new recyclable materials.
  
  Our Research Framework
- Research Framework
Services
- Royce brings together world-leading expertise and technical capabilities and works closely with industry to ensure translation and commercialisation of fundamental research.
  
  Our Services
Equipment
- Royce coordinates over 700 academic, technical, and research staff and over £200 million of facilities, providing a joined-up framework that can deliver beyond the current capabilities of individual Partners or research teams.
  
  Equipment & Facilities
News
Events
Contact

Home / Programmes / Digital Materials Foundry / Machine Learning for Property Prediction

Overview

This page catalogues a series of machine-learning models and source code for predicting properties from chemical compositions of functional materials.

All predictions employ computationally efficient machine-learning methods to aid sustainability needs for compute.

Library

The following resources are courtesy of: University of Cambridge Royce Partner. Digital library assembly and maintenance funded via the EPSRC AI Hubs, AIchemy [aichemy.org] (EP/Y028775/1, EP/Y028759/1), APRIL [april.ac.uk] (EP/Y029763/1) and Royce (EP/…)

Ferromagnetic Materials

User Input: chemical composition of chosen material

Model Output: Curie temperature prediction

Licence: MIT

ML Model & Source Code Citation

Superconducting Materials

User Input: chemical composition of chosen material

Model output: predicted critical temperature of superconductivity

Licence: MIT

ML Model & Source Code Citation

Semiconducting Materials

User Input: chemical composition of chosen material

Model Output: band-gap energy prediction

License: MIT

ML Model & Source Code Citation

High-Entropy Alloys

User Input: chemical composition of chosen material

Model Output: prediction of various properties that include: elastic constants, Young’s modulus, shear modulus, Wigner–Seitz radius, formation enthalpy, total energy, Zener anisotropy, Pugh ratio, Poisson ratio, universal anisotropy.

Licence: MIT

ML Model & Source Code Citation

Amorphous Metallic Alloys

User Input: chemical composition of chosen material

Model Output: prediction of their glass forming ability and the most significant material features that govern this property

License: MIT

ML Model & Source Code Citation

Engage with the Digital Materials Foundry

The Digital Materials Foundry is a new programme within the Henry Royce Institute designed to address challenges around AI in Materials Discovery, Characterisation and Application. If you would like to submit work to its libraries of Experimental Materials Data Repositories or Machine Learning for Property Prediction, please get in touch using the link below.

Contact Us