Deflated HeteroPCA: Overcoming the curse of ill-conditioning in heteroskedastic PCA
This addresses a critical issue in statistical estimation for applications like factor models and tensor PCA, offering improved robustness for researchers and practitioners dealing with heteroskedastic data.
The paper tackles the problem of estimating the column subspace of a low-rank matrix from contaminated data with heteroskedastic noise and unbalanced dimensionality, proposing Deflated-HeteroPCA to overcome performance degradation due to ill-conditioning, achieving near-optimal and condition-number-free statistical accuracy with theoretical guarantees.
This paper is concerned with estimating the column subspace of a low-rank matrix $\boldsymbol{X}^\star \in \mathbb{R}^{n_1\times n_2}$ from contaminated data. How to obtain optimal statistical accuracy while accommodating the widest range of signal-to-noise ratios (SNRs) becomes particularly challenging in the presence of heteroskedastic noise and unbalanced dimensionality (i.e., $n_2\gg n_1$). While the state-of-the-art algorithm $\textsf{HeteroPCA}$ emerges as a powerful solution for solving this problem, it suffers from "the curse of ill-conditioning," namely, its performance degrades as the condition number of $\boldsymbol{X}^\star$ grows. In order to overcome this critical issue without compromising the range of allowable SNRs, we propose a novel algorithm, called $\textsf{Deflated-HeteroPCA}$, that achieves near-optimal and condition-number-free theoretical guarantees in terms of both $\ell_2$ and $\ell_{2,\infty}$ statistical accuracy. The proposed algorithm divides the spectrum of $\boldsymbol{X}^\star$ into well-conditioned and mutually well-separated subblocks, and applies $\textsf{HeteroPCA}$ to conquer each subblock successively. Further, an application of our algorithm and theory to two canonical examples -- the factor model and tensor PCA -- leads to remarkable improvement for each application.