Unique sparse decomposition of low rank matrices
This addresses a fundamental problem in matrix decomposition for applications in areas like signal processing and machine learning, but it appears incremental as it builds on existing sparse and low-rank decomposition frameworks.
The paper tackles the problem of uniquely decomposing a low-rank matrix into a sparse representation, proving that this decomposition can be uniquely identified up to signed permutation, with theoretical guarantees that strict local solutions are close to the ground truth and can be recovered using a data-driven initialization and second-order descent algorithms.
The problem of finding the unique low dimensional decomposition of a given matrix has been a fundamental and recurrent problem in many areas. In this paper, we study the problem of seeking a unique decomposition of a low rank matrix $Y\in \mathbb{R}^{p\times n}$ that admits a sparse representation. Specifically, we consider $Y = A X\in \mathbb{R}^{p\times n}$ where the matrix $A\in \mathbb{R}^{p\times r}$ has full column rank, with $r < \min\{n,p\}$, and the matrix $X\in \mathbb{R}^{r\times n}$ is element-wise sparse. We prove that this sparse decomposition of $Y$ can be uniquely identified, up to some intrinsic signed permutation. Our approach relies on solving a nonconvex optimization problem constrained over the unit sphere. Our geometric analysis for the nonconvex optimization landscape shows that any {\em strict} local solution is close to the ground truth solution, and can be recovered by a simple data-driven initialization followed with any second order descent algorithm. At last, we corroborate these theoretical results with numerical experiments.