RT Book,_Section
T1 A New Approach for Sparse Matrix Classification Based on Deep Learning Techniques
A1 Pichel Campos, Juan Carlos
A1 Pateiro López, Beatriz
K1 Sparse matrix
K1 Classification
K1 Deep Learning
K1 CNN
K1 Performance
K1 Sparse matrices
K1 Training
K1 Computer architecture
K1 Convolution
K1 Kernel
K1 Convolutional neural networks
AB In this paper, a new methodology to select the best storage format for sparse matrices based on deep learning techniques is introduced. We focus on the selection of the proper format for the sparse matrix-vector multiplication (SpMV), which is one of the most important computational kernels in many scientific and engineering applications. Our approach considers the sparsity pattern of the matrices as an image, using the RGB channels to code several of the matrix properties. As a consequence, we generate image datasets that include enough information to successfully train a Convolutional Neural Network (CNN). Considering GPUs as target platforms, the trained CNN selects the best storage format 90.1% of the time, obtaining 99.4% of the highest SpMV performance among the tested formats
PB IEEE
SN 978-1-5386-8319-4
YR 2018
FD 2018
LK http://hdl.handle.net/10347/18645
UL http://hdl.handle.net/10347/18645
LA eng
NO J. C. Pichel and B. Pateiro-López (2018), A New Approach for Sparse Matrix Classification Based on Deep Learning Techniques, 2018 IEEE International Conference on Cluster Computing (CLUSTER), Belfast, pp. 46-54. doi: 10.1109/CLUSTER.2018.00017
NO This work has been supported by MINECO (TIN2014-54565-JIN and MTM2016-76969-P), Xunta de Galicia (ED431G/08) and European Regional Development Fund
DS Minerva
RD 24 abr 2026