A new perspective on low-rank optimization
This work provides a general method for improving relaxations in low-rank optimization, which is incremental as it extends existing perspective reformulation techniques to matrix settings.
The paper tackles the problem of obtaining strong convex relaxations for low-rank optimization problems by characterizing convex hulls using the matrix perspective function and orthogonal projection matrices, resulting in tractable semidefinite constraints for applications like reduced rank regression and non-negative matrix factorization.
A key question in many low-rank problems throughout optimization, machine learning, and statistics is to characterize the convex hulls of simple low-rank sets and judiciously apply these convex hulls to obtain strong yet computationally tractable convex relaxations. We invoke the matrix perspective function - the matrix analog of the perspective function - and characterize explicitly the convex hull of epigraphs of simple matrix convex functions under low-rank constraints. Further, we combine the matrix perspective function with orthogonal projection matrices-the matrix analog of binary variables which capture the row-space of a matrix-to develop a matrix perspective reformulation technique that reliably obtains strong relaxations for a variety of low-rank problems, including reduced rank regression, non-negative matrix factorization, and factor analysis. Moreover, we establish that these relaxations can be modeled via semidefinite constraints and thus optimized over tractably. The proposed approach parallels and generalizes the perspective reformulation technique in mixed-integer optimization and leads to new relaxations for a broad class of problems.