Fast Algorithms for Quaternion-Valued Convolutional Neural Networks
Close
Conţinutul numărului revistei
Articolul precedent
Articolul urmator
321 0
SM ISO690:2012
CARIOW, Aleksandr, CARIOWA, G.. Fast Algorithms for Quaternion-Valued Convolutional Neural Networks. In: IEEE Transactions on Neural Networks and Learning Systems, 2021, nr. 1(32), pp. 457-462. ISSN 2162-237X. DOI: https://doi.org/10.1109/TNNLS.2020.2979682
EXPORT metadate:
Google Scholar
Crossref
CERIF

DataCite
Dublin Core
IEEE Transactions on Neural Networks and Learning Systems
Numărul 1(32) / 2021 / ISSN 2162-237X /ISSNe 2162-2388

Fast Algorithms for Quaternion-Valued Convolutional Neural Networks

DOI: https://doi.org/10.1109/TNNLS.2020.2979682

Pag. 457-462

Cariow Aleksandr1, Cariowa G.2
 
1 Technical University of Moldova,
2 West Pomeranian University of Technology of Szczecin, Poland
 
Disponibil în IBN: 17 ianuarie 2021


Rezumat

In this article, we analyze algorithmic ways to reduce the arithmetic complexity of calculating quaternion-valued linear convolution and also synthesize a new algorithm for calculating this convolution. During the synthesis of the discussed algorithm, we use the fact that quaternion multiplication may be represented as a matrix-vector product. The matrix participating in the product has unique structural properties that allow performing its advantageous decomposition. Namely, this decomposition leads to reducing the multiplicative complexity of computations. In addition, we used the fact that when calculating the elements of the quaternion-valued convolution, the part of the calculations of all matrix-vector products is common. It gives an additional reduction in the number of additions of real numbers and, consequently, a decrease in the additive complexity of calculations. Thus, the use of the proposed algorithm will contribute to the acceleration of calculations in quaternion-valued convolution neural networks.

Cuvinte-cheie
Complex networks, Convolution, Matrix algebra, Parallel processing systems