Prospective students

Research Topics

I am looking for motivated students that want to do research with me on various privacy topics. My main research topics are:

• Privacy-preserving searchable encryption
• Machine learning privacy
• Anonymous communications systems
• Location privacy

Please check my Research Page for an informal description of each of my research lines (written, mostly, for prospective students). Besides these directions, I am open to work on privacy-related topics if there is common interest. Usually, if the problem involves reasoning about what privacy is, how to measure it, and how to provide it, I might be interested. My work sometimes uses cryptographic techniques, but I am not an expert (and thus not a good supervisor) for someone that wants to work in pure crypto, systems security, or network security. I do work on applied cryptography: develop systems that rely on available cryptographic tricks to reduce their leakage (e.g., privacy-preserving searchable databases).

What am I looking for?

Becoming a grad student is a big time commitment. You are going to spend at the very least (for master’s degrees) two years doing research on a topic. Being happy with your research topic (and with research itself) is very important! Do not do a research-based masters just for the sake of getting a master’s degree (we have a course-based master’s program if you do not want to do research). Thus, motivation is very important. Ask yourself questions like: why does research sound interesting to you? Why research on privacy? Why at UBC and/or with me?

Other skills that I am looking for:

• I usually think of data as a statistical process, and find statistical models to reason about privacy. Familiarity with statistical concepts such as Bayesian inference or estimation/detection is nice! You can also try reading a bit about differential privacy, and see if that is something that you’d feel comfortable working on.
• Privacy-preserving mechanisms can usually be designed by solving some sort of optimization problem. Familiarity with optimization techniques, or having done well in optimization courses, is useful. While machine learning is not required, I am open to using it in privacy design as well.
• If you’ve read cryptography papers, understood them, and found them interesting, that’s probably a good sign that working with me on something related to searchable encryption could be a good fit.
• For machine learning research: Proficiency in a machine learning library (I prefer PyTorch) is important. You should understand the basics of ML training (e.g., backpropagation), be familiar with basic architectures, and have performed well in ML courses during your degree.
• If the research we are going to do involves interacting with actual computer systems, getting measurements, and “getting hacky”, experience doing such things also helps. I am actually terrible with computers, but I’m happy to team up with someone that is skilled :)

Applying and Emails

I can only accept students that have applied through the general admission process of ECE UBC (MASc or PhD programs). Therefore, there is no need to send me an email letting me know you want to apply: I will review all the files of applicants that have selected my name as potential advisor, even if we haven’t talked prior. I get a lot of emails from prospective students, and sadly most of them are from students with loosely related research areas. I end up getting overwhelmed, putting all of them in a folder, and maybe read them when it’s too late.