Description
Instructor(s)/Supervisor(s)/Coordinator(s): Jize ZHANGThis course aims to introduce scientific machine learning to graduate students majoring in civil and environmental engineering. The course starts with an extensive review of probability and statistics, the difference between machine learning and descriptive statistics, discusses Monte Carlo sampling for uncertainty quantification and risk assessment, covers the basics of supervised (Bayesian linear regression, Gaussian processes, deep neural networks), unsupervised learning and dimensional reduction (k-means clustering, principal component analysis, Gaussian mixtures), state space models (Kalman filters, and if time permits, particle filters). The course will conclude with a look at the state-of-the-art in physics-informed machine learning and design-of-experiment for data-efficient modeling of civil infrastructure systems. By the end of this course, students will be able to identify the difference between a supervised (classification) and unsupervised (clustering, dimension reduction) technique, identify which technique they need to apply for a specific task or a particular dataset and need, engineer features to meet that need, and write Python-based code to carry out a machine learning analysis for civil infrastructure system modeling.