Stephan Rave (Uni Münster): Hierarchical Approximate SVD
Wednesday, 22.11.2023 14:15 im Raum M5
Truncated singular value decomposition (TSVD) is an important tool for extracting low-rank structure from high-dimensional data sets. In this talk, we present a simple approach to compute TSVDs of very large data sets in distributed, heterogeneous computing environments, by splitting the TSVD computation into a hierarchy of smaller sub-TSVDs. Rigorous bounds on both the approximation error and the efficiency of the approach are provided.
Angelegt am Wednesday, 16.08.2023 16:49 von Besprechungsraum
Geändert am Monday, 20.11.2023 13:59 von Stephan Rave
[Edit | Vorlage]