|About this Abstract
||2016 TMS Annual Meeting & Exhibition
||ICME Infrastructure Development for Accelerated Materials Design: Data Repositories, Informatics, and Computational Tools
||Towards Better Efficiency and Accuracy: Data Mining for Prediction and Optimization in Materials System Design
||Ankit Agrawal, Alok Choudhary
|On-Site Speaker (Planned)
Deciphering the processing-structure-properties-performance (PSPP) linkages in materials is at the heart of computational materials science. The application of high performance data mining techniques in materials science opens up new avenues for accelerated materials discovery and design, the need for which has also been emphasized by the Materials Genome Initiative. In this talk, I will describe some of the recent works done in our group in collaboration with several materials scientists, employing state-of-the-art data analytics for exploring PSPP linkages, both in terms of forward models (e.g. predicting a material property for a given composition and/or structure) and inverse models (e.g. discovering material compositions and/or structures that possess a desired property). Illustrative works include data-driven analytics on both simulation data like density functional theory (DFT), and experimental data such as processing and composition parameters of steels. Results indicate that such data-driven analytics can significantly accelerate the prediction/optimization process for materials design.
||Planned: A print-only volume