Machine Learning for Mechanical Engineering

ME 405 treats machine learning as a working tool for mechanical engineers, focused on turning experimental and simulation data — microscopy images, materials databases, atomistic descriptors — into models that predict structure-property relationships or drive a self-driving lab. You'll move through the full pipeline in Python: cleaning and curating data, fitting regression and classification models, then progressing to neural networks, Gaussian processes, and reinforcement learning, with two homeworks and a project where you apply this to a real ME problem. It assumes you're comfortable with programming and statistics, and it's the bridge between traditional ME coursework and the data-driven side of modern materials and mechanics research.