Sparse Bayesian learning for basis selection

By David Paul Wipf, Bhaskar D. Rao. IEEE Transactions on Signal Processing, 2004

1,548 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Networks and Communications, Signal Processing

View Paper: https://www.semanticscholar.org/paper/0ca26f9a98dda0abb737692f72ffa682df14cb2f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Networks and Communications

In Signal Processing

Citations in the corpus

Iterative Channel Estimation Using LSE and Sparse Message Passing for MmWave MIMO Systems (2016, methodology): “The algorithms in the third category do not need any priori knowledge of the channel distribution except the noise variance, e.g., Sparse Bayesian Learning (SBL) [30], LASSO [32], Expectation-maximization Bernoulli-Gaussian Approximate Message Passing (EM-BG-AMP) [33], etc.”
Optimization with Sparsity-Inducing Penalties (2011, methodology): “In fact, sparsity should be understood as the requirement or expectation that a few coefficients are significantly larger than most of the rest, an idea which is somehow formalized as compressibility in the compressed sensing literature, and which inspires automatic relevance determination methods (”
Expectation-Maximization Gaussian-Mixture Approximate Message Passing (2012, methodology): “versions, EM-BG-AMP, genie-tuned13 Orthogonal Matching Pursuit (OMP) [29], genie-tuned13 Subspace Pursuit (SP) [30], Bayesian Compressive Sensing (BCS) [19], Sparse Bayesian Learning [18] (via the more robust T-MSBL [31]), de-biased genie-tuned14 LASSO (via SPGL1 [32]), and Smoothed(SL0) [33].”
Sparse Signal Recovery With Temporally Correlated Source Vectors Using Sparse Bayesian Learning (2011, methodology): “Note that γ is a function of the estimated B i ( ∀ i ). b However, the theorem implies that even when the estimated B i is different from the true B i , the estimated sources are the true sources at the global minimum of the cost function.”
Identification of isolated structural damage from incomplete spectrum information using l1-norm minimization (2014, methodology): “After computing the posterior, identifies maximum a posteriori estimate or region with large probability mass [30].”
Compressed Sensing for Wireless Communications: Useful Tips and Tricks (2015, methodology): “It has been shown that with appropriately chosen parameters, the SBL algorithm outperforms 1 norm minimization and iteratively reweighted algorithms [98].”
Compressive sensing: From theory to applications, a survey (2013, methodology): “There are many algorithms proposed in literature that use this technique like focal underdetermined system solution (FOCUSS) [48], iterative re-weighted least squares [49], sparse bayesian learning algorithms [50], Monte-Carlo based algorithms [51], etc.”
AMP-Net: Denoising-Based Deep Unfolding for Compressive Image Sensing (2020, methodology): “A number of non-linear iterative algorithms can be used to solve these optimization problems [11], [21], [22], such as sparse Bayesian learning [23], orthogonal matching pursuit (OMP) [24], fast iterative shrinkage-thresholding algorithm (FISTA) [25], approximate message passing (AMP) [26], etc.”
Gaussian Kullback-Leibler approximate inference (2013, background): “Examples include, feature selection in regression problems (Wipf, 2004), source separation (Girolami, 2001), denoising or deblurring problems (Fergus et al.”
Modified-CS: Modifying compressive sensing for problems with partially known support (2009, background): “ other potential applications where sparse reconstruction for time sequences of signals/images may be needed, e.g. see [3], [4]. Sparse reconstruction has been well studied for a while, e.g. see [5], [6]. Recent work on Compressed Sensing (CS) gives conditions for its exact reconstruction [7], [8], ”
Sparse Coding via Thresholding and Local Competition in Neural Circuits (2008, methodology): “For example, it has been shown that both the spatial and temporal properties of V1 receptive ﬁelds may be accounted for in terms of a dictionary that has been optimized for sparseness in response to natural images (Olshausen, 2003).”
Gradient Pursuits (2008, neutral): “Nonconvex local optimization procedures include the focal underdetermined system solver (FOCUSS) [11] and Bayesian approaches, such as the relevance vector machine, also known as sparse Bayesian learning [12], [13] or Monte Carlo-based approaches, such as those in [14]–[18].”
Multitask Compressive Sensing (2009, methodology): “5A simple modification to (12) is available from [42] by exploiting the matrix inversion lemma, which leads to an operation per iteration.”
Sparse Representation for Signal Classification (2006, methodology): “Except for the intuitive interpretation as obtaining a sparse factorization that minimizes signal reconstruction error, the problem formulated in equation (3) has an equivalent interpretation in the framework of Bayesian decision as follows [13].”
Bayesian Compressive Sensing (2008, background): “4A simple modification to (10) is available from [25] by exploiting the matrix inverse identity, which leads to anO(K ) operation per iteration.”
Sparse Methods for Direction-of-Arrival Estimation (2016, background): “The first class is within the framework of SBL (see, e.g., [119–122]).”