1: \begin{abstract}

2: Sparse Bayesian learning (SBL) can be implemented with low complexity based on the approximate message passing (AMP) algorithm. However, it \rev{does not work well for a generic measurement matrix}, which may cause AMP to diverge. Damped AMP has been used for SBL to alleviate the problem at the cost of reducing convergence speed. In this work, we propose a new SBL algorithm based on structured variational inference, leveraging AMP with a unitary transformation (UAMP). Both single measurement vector and multiple measurement vector problems are investigated. It is shown that, compared to state-of-the-art AMP-based SBL algorithms, the proposed UAMP-SBL is more robust and efficient, leading to remarkably better performance. 

3: %\tcr{ and its consistency with state-evolution predictions.}

4: %In many cases UTAMP-SBL can approach the support-oracle bound closely.

5: %Furthermore, the UTAMP-SBL is also applied to dealing with the problem of sparse signal %recovery with multiple measurement vectors (MMV), where the sparse signal vectors share %a common support. Experimental results indicted that UTAMP-SBL still delivers %considerably better robustness and computation advantages than existing algorithms.

6:   

7: \end{abstract}

8: