Examining the intersection of materials science and machine learning

In July 2023, the Center of Excellence on Data-Driven Discovery of Multifunctional Materials Systems (D3OM2S) hosted a workshop on the current challenges and opportunities for machine learning in the context of material design and fabrication in Dayton, OH.

The workshop brought together researchers at all levels and featured contributors from academia, industry and government in order to identify future research topics where collaborations between Materials Science and Machine Learning have a unique advantage over separate efforts.

Over two days, a variety of topics were covered, including consensus equilibrium and steady state microstructure evolution, equivariance in physics informed networks involving crystallographic symmetry, decision science and the next most informative sequence of experiments, as well as large language models in materials synthesis and applications.

"It was exciting to learn about the advances in machine learning methods and their application to the development of advanced material systems," said professor Michael Bockstaller, who leads the D3OM2S initiative. "The workshop also presented unique opportunities for participants to network for the benefit both the research and teaching at Carnegie Mellon."

Participants in the workshop were able to learn about the fundamental and practical aspects of machine learning as they relate to applications in materials engineering. In addition to the topics outlined on the agenda, attendees had the opportunity to bring new ideas to the group in discussion sessions.