A non-backtracking method for long matrix completion

We consider the problem of rectangular matrix completion in the regime where the matrix $M$ of size $n\times m$ is ``long", i.e., the aspect ratio $m/n$ diverges to infinity. Such matrices are of particular interest in the study of tensor completion, where they arise from the unfolding of an odd-order low-rank tensor. In the case where the sampling probability is $\frac{d}{\sqrt{mn}}$, we propose a new algorithm for recovering the singular values and left singular vectors of the original matrix based on a variant of the standard non-backtracking operator of a suitably defined bipartite graph. We show that when $d$ is above a Kesten-Stigum-type sampling threshold, our algorithm recovers a correlated version of the singular value decomposition of $M$ with quantifiable error bounds. This is the first result in the regime of bounded $d$ for weak recovery and the first result for weak consistency when $d\to\infty$ arbitrarily slowly without any polylog factors.