Course talk:CPSC522/Differential Privacy

I put my critique in this PDF as comments. https://drive.google.com/file/d/1ln0RHuj8J2OyjYtUjCN27cKv6FlFOz30/view?usp=share_link

Thank you for your response! I will add the suggestions you have made.

The article is about DP, which I think is not super relevant to this course but obviously has applications in AI/ML. The mathematics could be explained better, rest all is good. It would have been better if there was a discussion of applications to AI/ML, current trends (like in federated learning), and uses in companies like Google as you mentioned in the starting. On similar grounds, recent papers could have been pointed to would be useful in related work.

Thank you for your response! Although DP is not directly related to CPSC 522, most AI and ML methods can be extended to DP, including my previous wiki page that privatizes MDP which is very relevant to this course. I would've loved to do an FL page but it is not very relevant to the course and is beyond the scope of this page. As for more up-to-date studies, I have included Renyi DP, as well as cited some other good reads that should be invaluable for anybody interested in DP.