MATH4365 - Introduction to Deep Learning

MATH 4365 Introduction to Deep Learning (3 semester credit hours) Topics include single and multilayer neural network models; loss and activation functions; backpropagation algorithm; common neural architectures for classification and regression; autoencoders; training deep neural networks; methods for improving generalizability of deep learners; recurrent and convolutional neural networks; and reinforcement learning. Computer packages such as R or Python will be used for implementation of methods and data analysis. Prerequisites: (STAT 4355 or STAT 4360) and (MATH 2418 or MATH 4355) or instructor consent required. (Same as STAT 4365) (3-0) Y