Mathematics
College of Science & Engineering
Twin Cities
Award Year
2024
University Award
McKnight Presidential Fellow Award
Robust and efficient algorithms for corrupted data
William Leeb researches the design and analysis of efficient algorithms that are robust to corruption of the input data. Sources of corruption can include missing values, geometric perturbations, noise, and other types of information loss. Datasets with these challenges are encountered across numerous domains, from structural biology to image processing. His research program includes the development of new tools from multiple branches of mathematics, including harmonic analysis and random matrix theory, and has led to fundamental contributions to the problems of matrix estimation, orbit recovery, and robust metric design.