1: \begin{abstract}

2: Phase retrieval problems involve solving linear equations, but with missing  sign (or phase, for complex numbers) information.

3: Over the last two decades, a popular generic empirical approach to the many variants of this problem has been one of alternating minimization;

4: i.e. alternating between estimating the missing phase information, and the candidate solution. In this paper, we show that a simple alternating

5: minimization algorithm geometrically converges to the solution of one such problem  -- finding a vector $\vecfont{x}$ from $\vecfont{y},\mat{A}$,

6: where $\vecfont{y} = |\mat{A}^T\vecfont{x}|$ and $|\vecfont{z}|$ denotes a vector of element-wise magnitudes of $\vecfont{z}$ -- under the assumption that $\mat{A}$ is Gaussian.

7: 

8: Empirically, our algorithm performs similar to recently proposed convex techniques for this variant (which are based on ``lifting" to a convex matrix problem) in sample complexity and robustness to noise.

9: However, our algorithm is much more efficient and can scale to large problems.

10: Analytically, we show geometric convergence to the solution, and sample complexity that is off by log factors from obvious lower bounds.

11: We also establish close to optimal scaling for the case when the unknown vector is sparse. Our work represents the only known

12: theoretical guarantee for alternating minimization for any variant of phase retrieval problems in the non-convex setting.

13: \end{abstract}

14: