Project proposal

Due date: October 7, 5pm (no late submissions allowed)

The goal of the project is to give you experience in research on topics in trustworthy ML. The project is intended to be a research project, and not just a survey of existing work. You are free to pursue a purely theoretical project, a purely empirical project, or some combination of these. For example, you could study a new application domain and investigate aspects of trustworthiness there, develop new algorithms, systematically evaluate and compare existing algorithms, show new theoretical guarantees, or propose new notions of trustworthiness. You can discuss project ideas with the instructors. The project will be in groups of two students.

The project proposal is worth 5% of the grade for the class. It should be about 1 page long with normal fonts and margins. It should cover the following:

• Title of your project, group members.
• A short abstract.
• A few paragraphs describing the motivation of the project, intended objectives and scope, and what makes the project interesting to you.
• A tentative timeline for the project, including milestones to achieve by the project presentation on 11/19, and the final report submission which will be due in the final exam period (between December 10 to 17, exact date to be announced later). Note that we do not expect your project to be complete by the presentation on 11/19, but you should have made some reasonable progress.

The paper will be due on Gradescope. Each team should make a single submission (make sure to mark your team member in the Gradescope submission). Please note that no late submissions are allowed, so plan accordingly.