Alex is a computer scientist whose research has focused on developing methods for statistical estimation and machine learning which provide formal privacy guarantees to individuals.
His PhD research at the University of Toronto, supervised by Aleksandar Nikolov and Toniann Pitassi, focused on factorization techniques under local differential privacy. Work with Nikolov and Ullman showed that factorization techniques could be adapted so as to be nearly optimal for solving a wide class of estimation problems under local differential privacy with nearly the fewest possible number of data points which each of those problems required. Work with Nikolov and Pitassi gave analogous results for learning, providing factorization techniques which they proved to be nearly optimal for a wide class of learning problems as well.
Alex continues to be interested in the technical aspects of privacy, while also exploring the practical dimensions of privacy technologies and how they interact with other concerns such as reproducibility, transparency, fairness, and copyright. Alex’s interests extend to AI safety more broadly.
Machine Learning; Privacy; Fairness; AI Safety
Areas of interest for Doctoral Supervision: Theoretical Computer Science; Learning Theory; Information Theory; Differential Privacy; AI