Course Information

Overview

We will study the foundations of trustworthy AI, including privacy, robustness, generalization, uncertainty quantification, alignment, copyright protection, and incentives, with the specific mix of topics shaped by students' interests. For each topic we will see a mix of more classical theoretical work and more recent research trying to apply these ideas to practice. The primarily deliverable will be a final course project and presentation.

Instructor

Jonathan Ullman

Office: 216 Mass Ave · #241
Office Hours: TBD
Email: j.ullman@northeastern.edu

Meetings

Time: TF 9:50–11:30am
Location: 296 Ryder Hall

Students are expected to attend and participate actively in all class meetings within reason. Occasional absenses due to illnesses or work obligations are inevitable and that is completely fine, but please notify me about more extended absenses. Meetings will be recorded and the recordings will be available on request, but these are not a substitute for attendance.

Final Project

The main course deliverable is a final project and a final in-class project presentation. The project can be on any topic related to the course content, and it is OK for your project to connect to your existing research agenda as long as the project itself is original work. Projects can be done solo or in a team of two, and groups of three or more may be allowed for especially ambitious projects. More information about project expectations and milestones will be shared in the first few weeks.