Privacy in the World of Big Data
Course Project Information
The goal of the course project is to deepen your understanding of how Computer Science techniques can be used to address privacy-related challenges and provide a framework for advancing the state-of-the-art. The project should strive to produce something novel such as:
- new algorithms for privacy-preserving data mining that outperform existing ones on some metric;
- an implementation or adaptation of existing work to a new domain together with experimental analyses of results demonstrating previously unknown aspects of it;
- a discovery and analysis of a new privacy vulnerability (talk to the instructor about ethics and IRB approval first);
- a new tool, technique or system for empowering individuals to protect their privacy in some context;
- a new measurement that provides insight into implications of current practices on privacy;
- a new synthesis of existing results or comparison of techniques that provides novel insights or perspectives.
In other words, if it's about (the science or engineering of) privacy, is interesting to you, and you are going to say or do something new, it's a legitimate project.
- Proposal (due March 7): a 2-3 page proposal that includes the problem definition and motivation for its importance, a brief survey of related work, a description of your planned approach and how it differs from past work, criteria for evaluation, and a list of milestones describing what you plan to accomplish by April 4th and May 2nd. The proposal should contain enough pointers to starter ideas, related work, data or discussion sources to convince me that your proposed work is interesting and feasible.
- Progress Report (due April 4): a 2-3 page progress report that describes your progress so far and remaining steps, any obstacles you have encountered and plans to overcome them, and contains a description of preliminary findings.
- Project Presentation (April 25): a 15-25 minute presentation (depending on the number of projects in the class) of the project and results, in a style of a conference talk.
- Final Project Report (due May 4th, 2pm): a 6-10 page conference-style report that includes an abstract summarizing your work, an introduction section motivating the problem, a related work section that differentiates your contributions, a section describing your system / algorithm / implementation / measurement methodology, a section on results, evaluation, or implication of findings, as appropriate, a brief future work section outlining remaining open questions, and a bibliography.
Projects can be done individually or in groups of two (with expectations of scope increasing accordingly). It will account for 55% of the total class grade: 15% proposal, 10% progress report, 10% presentation, 20% final report.
I encourage you to come talk to me about your ideas, progress and challenges throughout the semester.
A Google doc with project ideas: bit.ly/599PPIdeas