Applied Machine Learning for Materials Science and Engineering

Catalog description

This course provides students with modern machine learning methodologies applied to materials science and engineering data. Topics span regression, classification, clustering, dimensionality reduction, neural networks (including convolutional, graph, and generative architectures), inverse modeling, uncertainty estimation, and multi-task learning, with attention to feature engineering, hyperparameter optimization, and model validation. The course emphasizes interpretability and domain integration via using feature attribution, uncertainty-aware predictions, and hybrid physics-ML workflows applied to real materials datasets (e.g., microstructure images, spectroscopy, molecular graphs). Students will implement methods in Python, critically engage with recent literature, and complete an individual project applying interpretable ML tools to a materials science challenge.