| Scope |
Big data techniques are being adopted in materials science to sort and analyze large volumes of disparate data for scientific discovery. This informatics approach is particularly attractive for analyzing micrographs, which traditionally rely on qualitative observations. This symposium focuses on analyzing images or multi-dimensional data with data methods, including computer visualization, advanced analytics, machine learning, and digital image correlation, to identify physical descriptors and higher order relationships. A special emphasis will be on applying these techniques to improve our understanding of structure-property relationships.
Session topics include:
- Data mining and machine learning applied to atomic/mesoscale images and spectroscopic data to identify defects
- Informing processing methods like additive manufacturing
- Transfer learning from experimental data to models
- Correlating mechanical, electrical, and thermal properties with microstructures |