|
Abstract:
|
Data science is an increasingly important field in the technical world, and recently, materials scientists have begun to reap the benefits of data science tools such as machine learning. This was previously a nearly impossible task, as the databases for materials properties were small and scarce. Material properties of interest were either calculated by hand through complex mathematical relationships or simulated with powerful computing. However, now that materials databases are growing, materials scientists are utilizing data science tools to predict the extremes of materials properties, like stiffness and dielectric constant, and to fill in the gaps of materials databases. In an effort to introduce new materials science engineers to data science, our research group will design an interactive algorithm that will walk students through the process of machine learning using whilst predicting a material property of interest. The tool will be available publicly on Nanohub.org, a high-performance computing site run by Purdue University. The student will be able to see, run, and modify the Python code in a Jupyter notebook on Nanohub.org.
|
