ADVANCED MATERIALS & PROCESSES | NOVEMBER/DECEMBER 2025 10 MACHINE LEARNING | AI The MIT Schwarzman College of Computing and the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) launched the MIT–MBZUAI Collaborative Research Program. Under the five-year agreement, faculty, students, and staff will collaborate to advance the foundations of AI and its applications in scientific discovery, human thriving, and the health of the planet. mit.edu. BRIEF AlloyGPT OFFERS HELP FOR ALLOY DESIGN Researchers at Carnegie Mellon University, Pittsburgh, are developing the potential to train large language models (LLMs) to understand a novel alloy physics language in a similar manner to how ChatGPT works. Led by assistant professor Mohadeseh Taheri- Mousavi, the team developed AlloyGPT, which recognizes the relationship between composition, structure, and properties to generate novel designs for additively manufacturable (AM) structural alloys. The AlloyGPT model features dual functionality: It can accurately predict multiple phase structures and properties based on given alloy compositions, as well as suggest a comprehensive list of alloy compositions that meet stated design goals. “We have created an architecture that has learned the physics of alloys in order to design enhanced alloys that have the desired qualities for mechanical performance and manufacturability in a variety of applications,” says Taheri-Mousavi. The team built the autoregressive model by developing a language for the physics of alloys and training this generative AI model. Rather than analyzing words, the model examines compositions and structural features in a sentence format to understand how the composition, structure, and properties are connected. Unlike conventional iterative methods, which often face challenges in finding all possible solutions, AlloyGPT can provide a comprehensive list of elemental combinations to produce the desired materials properties requested. This is especially useful for designing gradient composition AM alloys in which gradual changes in materials properties exist across a single part. cmu.edu. AI LINKS ATOMIC STRUCTURE TO QUANTUM TECH Researchers at the DOE’s Oak Ridge National Laboratory, Tenn., and colleagues developed a method to determine the atomic A vacancy defect on europium zinc arsenide. Courtesy of Ganesh Narasimha/ORNL. origins of unusual material behavior. Their approach uses Bayesian deep learning, a form of artificial intelligence that combines probability theory and neural networks to analyze complex datasets with remarkable efficiency. The technique reduces the amount of time required for experiments by helping scientists explore sample regions widely and rapidly converge on important features that exhibit interesting properties. “This method makes it possible to study a material’s properties with much greater efficiency,” says ORNL’s Ganesh Narasimha. “Usually, we would need to scan a large region, and then several small regions, and perform spectro- scopy, which is very time consuming. Here, the AI algorithm takes control and does this process automatically and intelligently.” The study explored europium zinc arsenide, a magnetic semimetal known for its unique electronic behaviors. However, the method is generalizable across a wide variety of materials, say researchers. Using advanced scanning tunneling microscopy, the team discovered connections between atomic structures and electronic properties. They say their streamlined approach simplifies the discovery process and advances capabilities related to AI and quantum science. ornl.gov. Assistant professor Mohadeseh Taheri-Mousavi and postdoctoral researcher Bo Ni with their AlloyGPT model.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==