Block-Sparse Tensor Recovery

Liyang Lu,Zhaocheng Wang,Zhen Gao,Sheng Chen,H. Vincent Poor

DOI: https://doi.org/10.1109/TIT.2024.3447050

2024-08-24

Abstract:This work explores the fundamental problem of the recoverability of a sparse tensor being reconstructed from its compressed embodiment. We present a generalized model of block-sparse tensor recovery as a theoretical foundation, where concepts measuring holistic mutual incoherence property (MIP) of the measurement matrix set are defined. A representative algorithm based on the orthogonal matching pursuit (OMP) framework, called tensor generalized block OMP (T-GBOMP), is applied to the theoretical framework elaborated for analyzing both noiseless and noisy recovery conditions. Specifically, we present the exact recovery condition (ERC) and sufficient conditions for establishing it with consideration of different degrees of restriction. Reliable reconstruction conditions, in terms of the residual convergence, the estimated error and the signal-to-noise ratio bound, are established to reveal the computable theoretical interpretability based on the newly defined MIP, which we introduce. The flexibility of tensor recovery is highlighted, i.e., the reliable recovery can be guaranteed by optimizing MIP of the measurement matrix set. Analytical comparisons demonstrate that the theoretical results developed are tighter and less restrictive than the existing ones (if any). Further discussions provide tensor extensions for several classic greedy algorithms, indicating that the sophisticated results derived are universal and applicable to all these tensorized variants.

Signal Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to reliably recover a block - sparse tensor from linear measurements within the compressed sensing framework. Specifically, the researchers proposed a generalized block - sparse tensor recovery model and defined the overall mutual incoherence property (MIP) of the measurement matrix set to analyze the recovery conditions under noise - free and noisy conditions. The main contributions of the paper are as follows:

1. **Define a new MIP concept**: In order to theoretically explain the tensor recovery problem, the author defined a series of new MIP concepts, which can be used as indicators to measure whether the measurement matrix set is rich enough to allow reliable reconstruction. By adjusting the internal measurement matrices within the set, these newly - defined concepts can be optimized, thereby increasing the flexibility of tensor recovery.

2. **Construct a generalized block - sparse tensor recovery model**: The author proposed a new block structure, called shadow block sparsity, to quantify the number of correlated blocks between each measurement matrix and the non - zero support blocks of the sparse tensor. At the same time, the cases where the tensor has a regular or irregular block structure were considered, and the corresponding analytical solutions were given.

3. **Propose a generalized T - GBOMP algorithm**: Based on the orthogonal matching pursuit (OMP) framework, the author developed a generalized tensor generalized block OMP (T - GBOMP) algorithm that includes multiple tensor block selection mechanisms. This algorithm is used as an important tool for deriving reliable reconstruction conditions for block - sparse tensor recovery. In addition, several tensorized algorithm variants based on existing greedy algorithms were summarized, and it was pointed out that the derived theoretical results are universal and applicable.

4. **In - depth theoretical analysis**: Based on the above preparatory work, the author carried out a detailed theoretical analysis, including the exact recovery conditions (ERC) and reliable reconstruction conditions under noise - free and noisy conditions. Specifically:
- A more relaxed exact recovery condition and its sufficient condition for T - GBOMP were proposed, ensuring that when the sparsity is within a certain upper bound, T - GBOMP can exactly recover the sparsest tensor from the measurements.
- The exponential convergence of the residual tensor was proved, and the upper and lower bounds of the shadow block sparsity and their corresponding tensor structures were given.
- The noisy recovery conditions based on MIP were derived, including the error upper bound between the original tensor and the estimated tensor, the lower bound of the SNR required for reliable recovery, and the norm of the residual tensor at a specific number of iterations.

These theoretical results are not only more compact and less restrictive than the existing results, but also demonstrate the superiority of block - sparse tensor recovery in multiple methods.