Chapel Hill, North Carolina
› Investigated the extent to which information can be inferred about website visits in encrypted connections; designed and implemented a software library and set of tools for performing statistically robust experiments and analyses on data representing such connections; compared a new approach with more than 10 approaches used in prior work under a new adversarial scenario as well as those used by previous work.
› Explored the security, usability, and linguistic challenges of using pronounceable tokens, and particularly lexical blends (i.e., portmanteaus), as passwords; developed methods and software for rating the pronounceability of word-like strings.
› Designed and analyzed multiple techniques for fast identification of opaque, i.e., compressed or encrypted, network traffic; evaluated techniques using the Bro and Snort intrusion detection systems on two high-speed campus networks.
› Explored extent to which automated techniques can reconstruct typed input from compromising reflections captured by commodity video cameras.
› Investigated severity of information leaks in encrypted VoIP conversations; designed and implemented an extensible object-oriented platform for sequence classification, including an implementation of profile hidden Markov models and other machine learning algorithms.
› Analyzed domain-name registrations to assess extent of speculation, tasting and front-running; designed and implemented a distributed system for measuring front-running.