Abstract:This paper investigates the impact of loss function design in deep unfolding techniques for sparse signal recovery algorithms. We focus on deep unfolded versions of the fundamental iterative shrinkage thresholding algorithm (ISTA) and the iterative hard thresholding algorithm (IHT). To obtain a guideline for the loss function design, we examine the effect of supervised learning using mean squared error and unsupervised learning using the objective function of the original optimization problem. Our simulation results reveal that the effect of loss function design significantly depends on the convexity of the optimization problem. For convex $\ell_1$-regularized problems, supervised-ISTA achieves better final recovery accuracy but fails to minimize the original objective function, whereas we empirically observe that unsupervised-ISTA converges to a nearly identical solution as conventional ISTA but with accelerated convergence. Conversely, for nonconvex $\ell_0$-regularized problems, both supervised-IHT and unsupervised-IHT converge to better local minima than the original IHT, showing similar performance under the training conditions regardless of the loss function employed. However, when the test conditions differ from the training conditions, our results suggest that unsupervised learning offers better robustness to distribution mismatch. These findings provide valuable insights into the design of effective deep unfolded networks for sparse signal recovery applications.
From: Ryo Hayakawa [view email]
[v1]
Mon, 1 Sep 2025 10:13:56 UTC (1,045 KB)
[v2]
Thu, 23 Apr 2026 12:47:03 UTC (1,757 KB)
[v3]
Fri, 10 Jul 2026 12:01:39 UTC (1,758 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。