MDI faculty and students make impact at the Materials Research Society’s annual meeting

by Nicole Capozziello

Assistant professor Kristofer Reyes was the lead organizer of a week-long symposium on machine learning and data-driven materials development and design. The symposium explored the synthesis of machine learning with materials research, highlighting a broad spectrum of topics in which machine learning, artificial intelligence, and/or statistics play a significant role in addressing problems in experimental and theoretical materials science. It was one of the most heavily attended symposiums in MRS history.

“The entire materials community has acknowledged that this topic is one of considerable importance, and will ultimately reshape how materials research is done – and our department is at the forefront of this transformation,” says Reyes.

As a follow-up to the symposium, Reyes is serving on the MRS staging committee on Artificial Intelligence, which aims to steer the community’s activities on AI.

In addition to Reyes, Empire Innovation Professor and Erich Bloch Endowed Chair Krishna Rajan opened the symposium with his talk on “Informatics Driven Design of Chemically Complex Materials.” Rajan discussed how material informatics can transform the existing paradigm of accelerated materials discovery and multiscale design. He showed how data science methods can be used to discover new pathways for the chemical design of glasses and ceramics as well as uncover hidden information in the molecular scale structure of these and other amorphous solids.

Faculty members Olgo Wodo and Scott Broderick, and numerous students from the department, also presented during the week-long event.

Earlier in the week, Reyes presented a popular tutorial entitled “An Introduction to Machine Learning Methods for Materials Science.” He introduced a wide variety of machine learning topics that have found utility in real-world materials research. He reviewed fundamental topics in machine learning, such as supervised and unsupervised learning, reinforcement learning and Bayesian techniques and optimization. Practical tools and techniques for handling experimental data, in addition to extracting the relevant information from such data to make the applications of machine learning methods possible, were also covered.

“The full-day tutorial was essentially a condensed version of some of the courses we teach here in the MDI department,” added Reyes.

MDI doctoral students Aparajita Dasgupta, Yingje Gao and Thaicia Stona de Almeida won an award for Best Poster for their research on “Data Driven Analysis of Volcano Plots and Prediction of New Binary Catalysts.” The project aims to understand the effects of chemistry and composition on multicomponent alloy systems in functional materials like catalysts, perovskites, metallic glasses and structural alloys using machine learning techniques. Dasgupta attended the meeting and accepted the award. Several other graduate students from the department also presented posters at the meeting.

“The talent of our new faculty and students was recognized at the MRS meeting, further enhancing the Department of Materials Design and Innovation as the leader in the field of materials informatics,” said Rajan.

The meeting, which took place from Nov. 25-30, 2018, in Boston, is an opportunity for materials science professionals from industry and education to connect with and present research to an interdisciplinary, international audience.