[edit]
Construction of optimal spectral methods in phase retrieval
Proceedings of the 2nd Mathematical and Scientific Machine Learning Conference, PMLR 145:693-720, 2022.
Abstract
We consider the \emph{phase retrieval} problem, in which the observer wishes to recover a n-dimensional real or complex signal \bX⋆ from the (possibly noisy) observation of |\bPhi\bX⋆|, in which \bPhi is a matrix of size m×n. We consider a \emph{high-dimensional} setting where n,m→∞ with m/n=O(1), and a large class of (possibly correlated) random matrices \bPhi and observation channels. Spectral methods are a powerful tool to obtain approximate observations of the signal \bX⋆ which can be then used as initialization for a subsequent algorithm, at a low computational cost. In this paper, we extend and unify previous results and approaches on spectral methods for the phase retrieval problem. More precisely, we combine the linearization of message-passing algorithms and the analysis of the \emph{Bethe Hessian}, a classical tool of statistical physics. Using this toolbox, we show how to derive optimal spectral methods for arbitrary channel noise and right-unitarily invariant matrix \bPhi, in an automated manner (i.e. with no optimization over any hyperparameter or preprocessing function).