Sparsity-Aware Tensor Decomposition

Sparse tensor decomposition, such as Canonical Polyadic Decomposition (CPD), is a key operation for data analytics and machine learning. Its computation is dominated by a set of MTTKRP (Matricized Tensor Times Khatri Rao Product) operations. The collection of required MTTKRP operations for sparse CP...

Full description

Saved in:
Bibliographic Details
Published in2022 IEEE International Parallel and Distributed Processing Symposium (IPDPS) pp. 952 - 962
Main Authors Kurt, Sureyya Emre, Raje, Saurabh, Sukumaran-Rajam, Aravind, Sadayappan, P.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2022
Subjects
CPD
Data analysis
Data models
Distributed processing
Load management
Machine learning
MTTKRP
Solid modeling
sparse tensor factorization
Tensors
Online AccessGet full text

Cover

More Information
Summary:Sparse tensor decomposition, such as Canonical Polyadic Decomposition (CPD), is a key operation for data analytics and machine learning. Its computation is dominated by a set of MTTKRP (Matricized Tensor Times Khatri Rao Product) operations. The collection of required MTTKRP operations for sparse CPD include common sub-computations across them and many approaches exist to factorize and reuse common sub-expressions. Prior work on sparse CPD has focused on minimizing the number of high-level operators. In this paper, we consider a design space that covers whether the partial MTTKRP results should be saved, different mode permutations and model the total volume of data movement to/from memory. Also, we propose a fine-grained load balancing method that supports higher levels of parallelization.
ISSN:1530-2075
DOI:10.1109/IPDPS53621.2022.00097