The power of small initialization in noisy low-tubal-rank tensor recovery

TL;DR

This paper demonstrates that small initialization in factorized gradient descent significantly improves low-tubal-rank tensor recovery from noisy measurements, achieving near-optimal error bounds regardless of overestimated tubal-rank.

Contribution

It introduces a four-stage analysis showing small initialization enables nearly minimax optimal recovery error in noisy tensor recovery, independent of over-parameterization.

Findings

Small initialization improves recovery accuracy in noisy tensor problems.

Theoretical error bounds are independent of overestimated tubal-rank R.

Early stopping strategies can achieve optimal practical results.

Abstract

We study the problem of recovering a low-tubal-rank tensor $\mathcal{X}\_\star\in \mathbb{R}^{n \times n \times k}$ from noisy linear measurements under the t-product framework. A widely adopted strategy involves factorizing the optimization variable as $\mathcal{U} * \mathcal{U}^\top$, where $\mathcal{U} \in \mathbb{R}^{n \times R \times k}$, followed by applying factorized gradient descent (FGD) to solve the resulting optimization problem. Since the tubal-rank $r$ of the underlying tensor $\mathcal{X}_\star$ is typically unknown, this method often assumes $r < R \le n$, a regime known as over-parameterization. However, when the measurements are corrupted by some dense noise (e.g., Gaussian noise), FGD with the commonly used spectral initialization yields a recovery error that grows linearly with the over-estimated tubal-rank $R$. To address this issue, we show that using a small initialization enables FGD to achieve a nearly minimax optimal recovery error, even when the tubal-rank $R$ is significantly overestimated. Using a four-stage analytic framework, we analyze this phenomenon and establish the sharpest known error bound to date, which is independent of the overestimated tubal-rank $R$. Furthermore, we provide a theoretical guarantee showing that an easy-to-use early stopping strategy can achieve the best known result in practice. All these theoretical findings are validated through a series of simulations and real-data experiments.