Materials research and development is an immense and enormously important market across various fields of technology, including semiconductors, aerospace, and automotive. However, conventional materials research relies on an outdated trial-and-error approach, which is a very slow and costly process. At Radical AI, we are revolutionizing the way materials are discovered by building a continuous loop of computation and experimentation driven by artificial intelligence. Specifically, we train accurate and efficient atomistic foundation models on large DFT databases to enable molecular dynamics simulations at 70 million times the speed of DFT. Promising materials are then synthesized and characterized in our laboratories, in an automated and autonomous way. Feedback from computational and experimental data is leveraged by our generative AI models to accelerate the discovery of new materials. As an example, I'll focus here on the case of bulk metallic glasses and highlight the promise and challenges of this class of materials.
*This work was supported primarily by the NSF through the Harvard University Materials Research Science and Engineering Center Grant No. DMR-2011754, US Department of Energy, Office of Basic Energy Sciences Award No. DE-SC0022199 as well as by the Camille and Henry Dreyfus Foundation Grant No. ML-22-075, the Department of Navy award N00014-20-1-2418 issued by the Office of Naval Research and Robert Bosch LLC. S.F. was supported by the Swiss National Science Foundation through the Postdoc mobility fellowship under grant number P500PT_214445. C.J.O. was supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. (DGE1745303). A.J. was supported by Aker Scholarship.
–
Publication:S. Falletta, A. Cepellotti, A. Johansson, C. W. Tan, A. Musaelian, C. J. Owen, B. Kozinsky, arXiv:2403.17207 (2024)
