Sie können Bookmarks mittels Listen verwalten, loggen Sie sich dafür bitte in Ihr SLUB Benutzerkonto ein.
Medientyp:
E-Artikel
Titel:
Convergence of an alternating direction and projection method for sparse dictionary learning
Beteiligte:
Zhou, Yijia;
Xu, Lijun
Erschienen:
IOP Publishing, 2020
Erschienen in:
Journal of Physics: Conference Series, 1592 (2020) 1, Seite 012066
Sprache:
Nicht zu entscheiden
DOI:
10.1088/1742-6596/1592/1/012066
ISSN:
1742-6588;
1742-6596
Entstehung:
Anmerkungen:
Beschreibung:
<jats:title>Abstract</jats:title>
<jats:p>In this paper, we focus on sparse dictionary learning that is widely used as a data processing technique in many real-world applications. Based on the frame of alternating direction method of multiplier (ADMM), we extend to an alternating direction and projection method for sparse dictionary learning. By introducing proximal mapping and the equivalence to the corresponding projection, a partial convergence result of this multi-block and nonconvex ADMM algorithm is given that the algorithm converges to a Karush-Kuhn-Tucker point whenever it converges.</jats:p>