Exact Reconstruction of Euclidean Distance Geometry Problem Using Low-Rank Matrix Completion

DOI10.1109/TIT.2018.2881749zbMATH Open1432.90117arXiv1804.04310OpenAlexW2963286770WikidataQ128913755 ScholiaQ128913755MaRDI QIDQ5223989FDOQ5223989

Authors: Abiy Tasissa, Rongjie Lai 

Publication date: 19 July 2019

Published in: IEEE Transactions on Information Theory (Search for Journal in Brave)

Abstract: The Euclidean distance geometry problem arises in a wide variety of applications, from determining molecular conformations in computational chemistry to localization in sensor networks. When the distance information is incomplete, the problem can be formulated as a nuclear norm minimization problem. In this paper, this minimization program is recast as a matrix completion problem of a low-rank $r$ Gram matrix with respect to a suitable basis. The well known restricted isometry property can not be satisfied in this scenario. Instead, a dual basis approach is introduced to theoretically analyze the reconstruction problem. If the Gram matrix satisfies certain coherence conditions with parameter $u$, the main result shows that the underlying configuration of $n$ points can be recovered with very high probability from $O(nrulo{g}^2(n))$ uniformly random samples. Computationally, simple and fast algorithms are designed to solve the Euclidean distance geometry problem. Numerical tests on different three dimensional data and protein molecules validate effectiveness and efficiency of the proposed algorithms.

Full work available at URL: https://arxiv.org/abs/1804.04310

Mathematics Subject Classification ID

Convex programming (90C25) Matrix completion problems (15A83)

Cited In (8)

• On the uniqueness of Euclidean distance matrix completions.
• On density extrema for digital discs
• Low-rank factorization for rank minimization with nonconvex regularizers
• Phase retrieval of complex and vector-valued functions
• A dual basis approach to multidimensional scaling
• Low-rank matrix completion in a general non-orthogonal basis
• Central limit theorems for classical multidimensional scaling
• Noisy low-rank matrix completion under general bases

This page was built for publication: Exact Reconstruction of Euclidean Distance Geometry Problem Using Low-Rank Matrix Completion

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5223989)