Par-VSOM: Parallel and Stochastic Self-organizing Map Training Algorithm

Omar Rivera-Morales, Lutz Hamel

2022

Abstract

This work proposes Par-VSOM, a novel parallel version of VSOM, a very efficient implementation of stochastic training for self-organizing maps inspired by ideas from tensor algebra. The new algorithm is implemented using parallel kernels on GPU accelerators. It provides performance increases over the original VSOM algorithm, PyTorch Quicksom parallel version, Tensorflow Xpysom parallel variant, as well as Kohonen’s classic iterative implementation. Here we develop the algorithm in some detail and then demonstrate its performance on several real-world datasets. We also demonstrate that our new algorithm does not sacrifice map quality for speed using the convergence index quality assessment.

Paper Citation

in Harvard Style

Rivera-Morales O. and Hamel L. (2022). Par-VSOM: Parallel and Stochastic Self-organizing Map Training Algorithm. In Proceedings of the 14th International Joint Conference on Computational Intelligence (IJCCI 2022) - Volume 1: NCTA; ISBN 978-989-758-611-8, SciTePress, pages 339-348. DOI: 10.5220/0011377700003332

in Bibtex Style

@conference{ncta22,
author={Omar Rivera-Morales and Lutz Hamel},
title={Par-VSOM: Parallel and Stochastic Self-organizing Map Training Algorithm},
booktitle={Proceedings of the 14th International Joint Conference on Computational Intelligence (IJCCI 2022) - Volume 1: NCTA},
year={2022},
pages={339-348},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011377700003332},
isbn={978-989-758-611-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Joint Conference on Computational Intelligence (IJCCI 2022) - Volume 1: NCTA
TI - Par-VSOM: Parallel and Stochastic Self-organizing Map Training Algorithm
SN - 978-989-758-611-8
AU - Rivera-Morales O.
AU - Hamel L.
PY - 2022
SP - 339
EP - 348
DO - 10.5220/0011377700003332
PB - SciTePress