Maximum diversity in web application user action privacy
R. Bagai, S. Layton, S. Gampa and K. Chandrashekar, "Maximum Diversity in Web Application User Action Privacy," 2019 IEEE 10th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON), New York City, NY, USA, 2019, pp. 0723-0729
Popular web application features, like autosuggestion and auto-correction, are well-known to amply reveal input actions entered by their users to any eavesdroppers, by generating sufficiently unique network activity bursts on actions. Some techniques, like k-anonymity and l-diversity, already exist in the literature for achieving different levels of privacy of user actions by making the generated network bursts for different actions indistinguishable from each other. We present a method that blends network bursts of not entire actions, but their probabilistic portions, with each other. Our method thus results in the maximum possible amount of l-diversity, namely k-diversity, and is a sianificant improvement over the existing ones.