Final Project Guidelines

CS229r: Mathematical Approaches to Data Privacy. Spring 2013

The final projects in the course are meant to give you the opportunity to further explore an aspect of differential privacy that interests you, give you the experience of formulating, carrying out, and presenting an interesting, short-term independent research project. We don't expect you to produce publishable results during time you have this semester, but we do hope that at least a few projects in the class can subsequently be developed into publishable research. Projects can be done individually or in pairs, with groups of three allowed for ambitious projects.

We are maintaining a list of topics to help inspire you on the course website, but you are free to come up with something from outside the list. Many different types of projects are possible:

• Experimental evaluation of algorithms or attacks on real-life datasets
• New theoretical results on some aspect of differential privacy or related topics
• Connecting differential privacy to some other area of interest to you (whether in CS or outside)
• Exploration of how differential privacy might be incorporated into a larger system design (with some particular application domain in mind)
• Modelling of some new problem that might not be captured by the current literature
• A synthesis and exposition of several papers in the differential privacy research literature (beyond those covered in class)
• And more... be creative!

The deadlines for various aspects of the project are as follows:

• Topic Ideas (Wednesday 3/13): Submit about a half-page of discussion about 1-3 ideas you have for potential project topics. For each topic, you should include the general question(s) that you'd like to address in your project, the general methodology (e.g. is it a theory project or an experimental project or...), and a list of papers in the research literature that seem relevant and which you plan to read. The point of these topic ideas is both to get you thinking about the project early on and to enable us to give you early feedback and suggestions, before you are too far into your project. If you already know who you plan to work with on the project, you may turn in a single set of topic ideas.
• Project Description (Friday 4/12): Submit a couple of pages giving a detailed description of what your final project will look like. You should be able to clearly state your research questions, briefly articulate how your project relates to what has been done in the past, describe the approach you are taking, give your timeline for completing various aspects of the project, and discuss your fallback plan in case you don't obtain the results that you're hoping to obtain.
• Written Paper (Friday 5/3): Submit a paper (approx. 10 pages) describing your completed project. The paper should motivate your research questions and results, explain how the project fits into the context of previous work, justify the methodology, and present the results in a convincing manner. (Naturally, the form will differ depending on the type of project.)
• Presentation (Tuesday 5/7 and Thursday 5/9): Every group will give an approx. 15 min presentation on their final project. Depending on the number of projects, we may need to start our class earlier than usual, so please keep 9-11:30 free on both days.

We encourage you to discuss your ideas with us throughout the process, at office hours or by email. We hope you enjoy the experience!