Matrix Recipes for Hard Thresholding Methods
Abstract
Given a set of possibly corrupted and incomplete linear measurements, we leverage low-dimensional models to best explain the data for provable solution quality in inversion. A non-exhaustive list of examples includes sparse vector and low-rank matrix approximation. Most of the well-known low dimensional models are inherently non-convex. However, recent approaches prefer convex surrogates that "relax" the problem in order to establish solution uniqueness and stability. In this paper, we tackle the linear inverse problems revolving around low-rank matrices by preserving their non-convex structure. To this end, we present and analyze a new set of sparse and low-rank recovery algorithms within the class of hard thresholding methods. We provide strategies on how to set up these algorithms via basic "ingredients" for different configurations to achieve complexity vs. accuracy tradeoffs. Moreover, we propose acceleration schemes by utilizing memory-based techniques and randomized, epsilon-approximate, low-rank projections to speed-up the convergence as well as decrease the computational costs in the recovery process. For all these cases, we present theoretical analysis that guarantees convergence under mild problem conditions. Simulation results demonstrate notable performance improvements compared to state-of-the-art algorithms both in terms of data reconstruction and computational complexity.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 07, 2012
- Accession Number
- ADA585818
Entities
People
- Anastasios Kyrillidis
- Volkan Cevher
Organizations
- Rice University