Deep Learning

Catalog description

Offered spring. Prereq., CSCI 547 . This course is a survey of state-of-the-art methods in deep learning. Covered topics are expected to change, but will typically include convolutional and recurrent neural networks, transformers, automatic differentiation, uncertainty quantification, generative models, deep reinforcement learning, training parallelism, model distillation, and interesting applications of all of the above to scientific problems. Students will read primary literature, present it to their peers, and develop computational implementations. Level: Graduate