- Researchers from Ames Laboratory and Texas A&M University are using AI to build machine learning models for the purpose of discovering and assessing compounds.
- Rare earth elements have a broad range of uses including energy storage and the development of clean energy technology.
- The project supervisor is optimistic that the machine learning model will find applications outside of rare-earth compounds research.
Researchers from Ames Laboratory and Texas A&M University are using artificial intelligence to discover and assess the stability of rare earth elements. The research is being performed through the Laboratory Directed Research and Development Program (LDRD) at Ames Laboratory.
Ames Laboratory has been at the forefront of rare-earths research since its beginning in 1947. Rare earth elements have a broad range of uses including energy storage and the development of clean energy technology.
ScienceDaily reports that the researchers are utilizing machine learning, a form of AI driven by algorithms and data usage, to discover and assess compounds. An advanced version of the Ames Laboratory Rare Earth database (RIC 2.0) was used alongside the high-throughput density-functional theory (DFT) to build the foundation for the machine learning model.
Recommended for You
High-throughput screening is a computational technique which enables researchers to rapidly test hundreds of models. On the other hand, DFT is a quantum mechanical method used to explore the thermodynamic and electronic properties of many body systems.
Yaroslav Mudryk, the project supervisor, explained that the machine learning model was designed primarily for the purpose of exploring rare earth compounds, but its application is not limited to rare-earths research. "It's not really meant to discover a particular compound," Mudryk stated. "It was, how do we design a new approach or a new tool for discovery and prediction of rare earth compounds? And that's what we did."
Mudryk highlighted that the team is exploring the machine learning model’s full potential, and are optimistic that a wide range of applications for the model will be available in the future.