Professor of Computer Science, Carla Gomes and her team of researchers and collaborators at the Institute for Computational Sustainability at Cornell (for which she serves as director and lead PI), are using AI and machine learning (ML) to make significant advances beyond the current reliance on hydrogen for fuel cells, since hydrogen is difficult to store. The team is searching for a catalyst for methanol to make fuel cells more efficient. However, as Melanie Lefkowitz writes in her article for the Cornell Chronicle, “no known materials are efficient catalysts for methanol oxidation,” and consequently, “a new material is needed.” As John Gregoire, Cornell CS Ph.D. ’09, currently at California Institute of Technology Joint Center for Artificial Photosynthesis, puts it: “If a viable catalyst exists, it’s going to need to be discovered by combining elements of the periodic table, and the number of combinations is so vast that it can’t be done with traditional experimentation.” Hence the need for machine learning to identify the “crystal structure” (or phase) of the material.

To mobilize AI and machine learning, Gomes and fellow researchers (including several current Cornell CS doctoral students—Sebastian Ament, Junwen Bai, Yiwei Bai, Johan Bj√∂rck, Di Chen, Brendan Rappazzo, M. Eng ’18, Wenting Zhao; Cornell CS postdoctoral researcher, Shufeng Kong; Yexiang Xue, Ph.D. ’18, currently Assistant Professor in computer science at Purdue University; Cornell CS programmer and analyst, Richard Bernstein; and contributions from Santosh K. Suram of Caltech, currently at the Toyota Research Institute) are developing innovative techniques combining AI reasoning and machine learning, including a system called CRYSTAL, which can aid in the mapping of crystal structures, “in which multiple bots each take on a different part of the problem, from predicting the phase structures of various combinations to making sure those predictions obey the rules of thermodynamics.” Details of findings can be discovered in a recent articlepublished in Materials Research Society (MRS) Communications (“CRYSTAL: A Multi-Agent AI System for Automated Mapping of Materials’ Crystal Structures”), in which:

We introduce CRYSTAL, a multi-agent AI system for crystal-structure phase mapping. CRYSTAL is the first system that can automatically generate a portfolio of physically meaningful phase diagrams for expert-user exploration and selection. CRYSTAL outperforms previous methods to solve the example Pd-Rh-Ta phase diagram, enabling the discovery of a mixed-intermetallic methanol oxidation electrocatalyst. The integration of multiple data-knowledge sources and learning and reasoning algorithms, combined with the exploitation of problem decompositions, relaxations, and parallelism, empowers AI to supersede human scientific data interpretation capabilities and enable otherwise inaccessible scientific discovery in materials science and beyond.

Results of the deployed multi-agent AI have been extremely effective. Already, researchers using CRYSTAL “were able to identify a unique catalyst, composed of three elements crystallized into a certain structure, that is effective for methanol oxidation and could be incorporated into methanol-based fuel cells.”

At a time when energy retention, production, and distribution is at the forefront of the necessary transition to renewable energy sources, innovations in fuel cell technology can yield a pronounced global impact. Not surprisingly, Gomes’ research team has attracted significant external funding, including from the National Science Foundation, the Army Research Office, the Air Force Office of Scientific Research, and the Toyota Research Institute among others. And for collaboration closer to Gates Hall, the researchers benefitted from the Cornell High Energy Synchroton Source.

Again, see the article in the Cornell Chronicle for additional remarks on the advances being made by Carla Gomes and her team of researchers and collaborators.