ADVANCED MATERIALS & PROCESSES | MAY 2026 10 MACHINE LEARNING | AI The Venado supercomputer at Los Alamos National Laboratory is now running OpenAI’s latest o3 reasoning model to accelerate national security research. Venado, which uses NVIDIA GH200 Grace Hopper Superchips, serves as a shared resource for researchers at the National Nuclear Security Administration laboratories. lanl.gov. BRIEF AI EXAMINES METAL 3D-PRINTED PARTS Scientists at the Korea Institute of Materials Science in collaboration with colleagues at the Max Planck Institute in Germany developed an artificial intelligence (AI)-based model capable of assessing the characteristics of internal defects during process design for metal additive manufacturing (AM). Conventional quality evaluation primarily focuses on simple indicators such as porosity. Yet the impact on mechanical performance varies sig- nificantly depending on the shape, size, location, and distribution of defects. To address these challenges, the team developed an explainable AI model capable of systematically analyzing and predicting the relationships among metal AM process conditions, defect morphology, and mechanical performance. This approach enables prediction of potential internal defects and their impact on performance. The core feature of the new AI model is its ability to analyze and predict internal defects generated during the laser powder bed fusion process of metal AM based on morphological characteristics rather than simply the number or fraction of defects. By using microstructural images, the model automatically analyzes pore size, non- circularity, and spatial distribution, and directly correlates these factors with mechanical properties, enabling a quantitative explanation of how defects influence actual performance. The team analyzed process conditions, powder characteristics, defect images, and mechanical property data across various metal AM materials, including steel, aluminum alloys, and titanium alloys, and used these datasets to train the AI model. Through this approach, they established an integrated framework capable of assessing how process variables and powder characteristics influence defect formation, and how this subsequently affects mechanical performance. www.kims.re.kr. TOOLKIT TURNS MICROSCOPY IMAGES INTO REAL DATA Researchers from The Hong Kong University of Science and Technology created an AI-enabled toolkit that automatically extracts and quantifies multiple microstructural features from microscopy images. Designed to meet the growing need for data-driven, autonomous workflows in materials science, the tool provides a systematic Graphical abstract. Courtesy of Acta Mater., 2026, doi.org/ 10.1016/j.actamat.2025.121751. method for converting complex image information into quantitative data. While modern microscopy can capture highly detailed images, the information they contain is often difficult to analyze consistently and at scale. Existing approaches typically focus on identifying simple features or classifying images. This limitation hinders the ability to fully understand structure-property re- lationships and slows down the design and optimization of new materials. To bridge this gap, the team designed GrainBot, which provides an integrated solution for segmentation, feature measurement, and correlation analysis. Using a convolutional neural network for precise grain segmentation, the toolkit is complemented by custom algorithms that can measure grain surface area, grain-boundary groove geometry, and surface concavity or convexity volumes. By converting each microscopy image into a rich set of numerical descriptors, GrainBot empowers researchers to build large-scale, standardized microstructure databases rather than relying on qualitative observations alone. www.hkust.edu.hk. Diagram illustrates the workflow of GrainBot, providing an integrated solution for segmentation, feature measurement, and correlation analysis.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==