This course will give introductions to research-related topics in mathematics. The topics will vary from semester to semester.

This course will survey developments in deep learning from the past 15 years, for a mathematically sophisticated audience with relatively little computer science background. Topics will include: - statistical inference and information theory. - neural nets and stochastic gradient descent. - convolutional neural networks. - reinforcement learning. - transformers and generative pre-training. The lecture emphasis will be on mathematical theory rather than implementation. However, enrolled students will be required to deliver in-class presentations and work on a final research project which involves implementation.

Pre-requisites: advanced mathematical background including analysis, abstract linear algebra, and probability theory; basic knowledge of data structures and algorithms; programming in Python.

Repeat Rules