Material Spatial Intelligence is a new paradigm in materials science, created by our reserach group, that leverages high-resolution spatial data of microstructure and local behavior to enable rapid and accurate prediction of macroscopic material properties. Unlike traditional approaches, which rely on bulk averages such as chemical composition or grain size, Material Spatial Intelligence encodes maps of the material’s microstructure and local behavior such as high resolution plastic field as into machine learning–ready representations.
By capturing the metals microstructural and behavioral heterogeneity using spatial data, this approach achieves unprecedented accuracy in property prediction and opens pathways for data-driven alloy design and qualification.
The central objective is to harness high-resolution spatial data, such as microstructural maps, strain fields, and defect distributions, and translate them into compact, machine-learning–ready representations. By doing so, we enable accurate prediction of key mechanical properties across diverse alloy systems, from steels to superalloys and additive manufactured metals. Beyond predictive accuracy, the project also builds generalizable workflows for spatial data integration and to develo open, shareable databases that accelerate discovery, design, and qualification of advanced materials.
The broader impact of this effort lies in its potential to transform materials development and qualification. By drastically reducing the time and cost associated with traditional testing campaigns, Material Spatial Intelligence can accelerate the deployment of advanced alloys in aerospace, automotive, energy, and space applications. The creation of open-access databases and AI-driven design tools will also lower entry barriers for researchers and industry, fostering innovation, education, and workforce training in data-driven materials engineering.
Selected Publications:
- NPJ Computational Materials : Learning Metal Microstructural Heterogeneity through Spatial Mapping of Diffraction Latent Space Features
- Scripta Materelia: Plasticity Encoding and Mapping during Elementary Loading for Accelerated Mechanical Properties Prediction
