A Machine Learning Approach for Identifying Defects in Amorphous Structure

Monday, September 13, 2021: 11:20 AM
227 (America's Center)
Ms. Tina Mirzaei , University of California, Riverside, Riverside, CA
We try to identify the defects by applying machine learning techniques to search through huge data sets of configurations in amorphous structures.We define defects as structures with local properties that are outliers in the property distribution and represent sites of weakness and can limit the strength of amorphous materials. For quantifying local structure, principal component analysis and cluster analysis will be presented to identify defect types. By including attributes that are known to be correlated with the property of interest as an input into the model, it becomes easier for a machine learning algorithm to automatically recognize these correlations and, thereby, create a more predictive model by analyzing large-scale data sets of configurations.