LSALSA: Accelerated Source Separation via Learned Sparse Coding

TL;DR

LSALSA introduces a learned deep architecture that accelerates sparse coding and source separation tasks, improving inference speed and accuracy by approximating an optimized ADMM-based algorithm with a trainable, unfolded network.

Contribution

The paper presents LSALSA, a novel deep learning framework that unrolls and learns parameters of SALSA, significantly speeding up sparse coding and source separation with theoretical analysis of its optimization properties.

Findings

Achieves faster inference times compared to traditional methods.

Improves quality of sparse code estimation and visual clarity.

Provides a theoretical understanding of the learned acceleration mechanism.

Abstract

We propose an efficient algorithm for the generalized sparse coding (SC) inference problem. The proposed framework applies to both the single dictionary setting, where each data point is represented as a sparse combination of the columns of one dictionary matrix, as well as the multiple dictionary setting as given in morphological component analysis (MCA), where the goal is to separate a signal into additive parts such that each part has distinct sparse representation within a corresponding dictionary. Both the SC task and its generalization via MCA have been cast as $\ell_1$-regularized least-squares optimization problems. To accelerate traditional acquisition of sparse codes, we propose a deep learning architecture that constitutes a trainable time-unfolded version of the Split Augmented Lagrangian Shrinkage Algorithm (SALSA), a special case of the Alternating Direction Method of Multipliers (ADMM). We empirically validate both variants of the algorithm, that we refer to as LSALSA (learned-SALSA), on image vision tasks and demonstrate that at inference our networks achieve vast improvements in terms of the running time, the quality of estimated sparse codes, and visual clarity on both classic SC and MCA problems. Finally, we present a theoretical framework for analyzing LSALSA network: we show that the proposed approach exactly implements a truncated ADMM applied to a new, learned cost function with curvature modified by one of the learned parameterized matrices. We extend a very recent Stochastic Alternating Optimization analysis framework to show that a gradient descent step along this learned loss landscape is equivalent to a modified gradient descent step along the original loss landscape. In this framework, the acceleration achieved by LSALSA could potentially be explained by the network's ability to learn a correction to the gradient direction of steeper descent.