Current browse context:
math.OC
Change to browse by:
References & Citations
Mathematics > Optimization and Control
Title: Weak Recovery Conditions from Graph Partitioning Bounds and Order Statistics
(Submitted on 28 Apr 2010 (v1), last revised 7 Nov 2012 (this version, v4))
Abstract: We study a weaker formulation of the nullspace property which guarantees recovery of sparse signals from linear measurements by l_1 minimization. We require this condition to hold only with high probability, given a distribution on the nullspace of the coding matrix A. Under some assumptions on the distribution of the reconstruction error, we show that testing these weak conditions means bounding the optimal value of two classical graph partitioning problems: the k-Dense-Subgraph and MaxCut problems. Both problems admit efficient, relatively tight relaxations and we use a randomization argument to produce new approximation bounds for k-Dense-Subgraph. We test the performance of our results on several families of coding matrices.
Submission history
From: Alexandre d'Aspremont [view email][v1] Wed, 28 Apr 2010 21:24:05 GMT (26kb)
[v2] Wed, 23 Jun 2010 00:36:51 GMT (27kb)
[v3] Tue, 28 Jun 2011 00:10:23 GMT (30kb)
[v4] Wed, 7 Nov 2012 13:34:02 GMT (35kb)
Link back to: arXiv, form interface, contact.