CHEM 4354 Machine Learning for Natural Sciences
This course explores the application of classical supervised and unsupervised learning methods, such as logistic regression, kernel method, Boltzmann machine, principal component analysis, autoencoder, and convolutional neural network, in various areas of natural sciences. Topics may include unsupervised detection of phase transition, dimension reduction and order parameter extraction, intermolecular forces fitting, graph neural networks for molecular design, trajectory prediction with time series analysis, deep learning analysis of images, protein folding, etc.
Cross Listed Courses
Due to the course's interdisciplinary nature, is also to be considered to satisfy a requirement fo the Physics minor as identified in the Minor Degree Plan for a Physics requirement.