An Asynchronous Parallel Approach to Sparse Recovery
This work addresses a specific bottleneck in parallel optimization for sparse recovery, offering an incremental improvement for researchers in signal processing and machine learning.
The paper tackles the challenge of applying asynchronous parallel computing to sparse recovery problems where the signal is sparse but cost functions are dense, proposing a stochastic greedy algorithm with asynchronous updates. Numerical simulations demonstrate potential benefits, though no specific performance numbers are provided.
Asynchronous parallel computing and sparse recovery are two areas that have received recent interest. Asynchronous algorithms are often studied to solve optimization problems where the cost function takes the form $\sum_{i=1}^M f_i(x)$, with a common assumption that each $f_i$ is sparse; that is, each $f_i$ acts only on a small number of components of $x\in\mathbb{R}^n$. Sparse recovery problems, such as compressed sensing, can be formulated as optimization problems, however, the cost functions $f_i$ are dense with respect to the components of $x$, and instead the signal $x$ is assumed to be sparse, meaning that it has only $s$ non-zeros where $s\ll n$. Here we address how one may use an asynchronous parallel architecture when the cost functions $f_i$ are not sparse in $x$, but rather the signal $x$ is sparse. We propose an asynchronous parallel approach to sparse recovery via a stochastic greedy algorithm, where multiple processors asynchronously update a vector in shared memory containing information on the estimated signal support. We include numerical simulations that illustrate the potential benefits of our proposed asynchronous method.