Matrix Denoising with Weighted Loss
This talk will describe a new class of methods for estimating a low-rank matrix from a noisy observed matrix, where the error is measured by a type of weighted loss function. Such loss functions arise naturally in a variety of problems, such as submatrix denoising, filtering heteroscedastic noise, and estimation with missing data. We introduce a family of spectral denoisers, which preserve the left and right singular subspaces of the observed matrix. Using new asymptotic results on the spiked covariance model in high dimensions, we derive the optimal spectral denoiser for weighted loss. We demonstrate the behavior of our method through numerical simulations.
William Leeb is an Assistant Professor in the School of Mathematics at the University of Minnesota, Twin Cities. He earned his PhD from Yale University in 2015 under the supervision of Ronald Coifman, and from 2015 to 2018 was a postdoc in Amit Singer's research group at Princeton University. William's research is in applied and computational harmonic analysis, statistical signal processing, and machine learning. He is particularly interested in estimation problems with low signal-to-noise ratios, high dimensionality, and many nuisance parameters.