BEGIN:VCALENDAR
PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN
VERSION:1.0
BEGIN:VEVENT
DTSTART:20151119T163000Z
DTEND:20151119T170000Z
LOCATION:18CD
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: For data analysis, a partial singular value decomposition (SVD) of the sparse matrix representing the data is a powerful tool. However, computing the SVD of a large data can take a significant amount of time even on a large-scale computer. Hence, there is a growing demand for a novel algorithm that can efficiently process the massive data being generated from many modern applications. To address this challenge, in this paper, we study randomized algorithms to update the SVD as changes are made to the data. Our experimental results demonstrate that these randomized algorithms can obtain the desired accuracy of the SVD with a small number of data accesses, and compared to the state-of-the-art updating algorithm, they often require much lower computational and communication costs. Our performance results on a hybrid CPU/GPU cluster show that these randomized algorithms can obtain significant speedups over the state-of-the-art updating algorithm.
SUMMARY:Randomized Algorithm to Update Partial Singular Value Decomposition on a Hybrid CPU/GPU Cluster
PRIORITY:3
END:VEVENT
END:VCALENDAR