High-Performance ELM for Memory Constrained Edge Computing Devices with Metal Performance Shaders

Anton Akusok, Leonardo Espinosa Leal, Kaj-Mikael Björk, Amaury Lendasse

Research output: Chapter in Book/Report/Conference proceedingChapterScientificpeer-review

Abstract

This paper proposes a block solution method for the Extreme Learning Machine. It combines the speed of a direct non-iterative solver with minimal memory requirements. The method is suitable for edge computing scenarios running on a mobile device with GPU acceleration. The implementation tested on the GPU of iPad Pro outperforms a laptop CPU, and trains a 19,000-neuron model using under one gigabyte of memory. It confirms the feasibility of Big Data analysis on modern mobile devices.
Original languageEnglish
Title of host publicationProceedings of ELM2019
EditorsJiuwen Cao, Chi Man Vong, Yoan Miche, Amaury Lendasse
PublisherSpringer
Publication date12.09.2020
Pages79-88
ISBN (Print)978-3-030-58988-2
ISBN (Electronic)978-3-030-58989-9
DOIs
Publication statusPublished - 12.09.2020
MoE publication typeA3 Book chapter

Publication series

NameProceedings of ELM2019
Volume14
ISSN (Print)2363-6084
ISSN (Electronic)2363-6092

Fingerprint Dive into the research topics of 'High-Performance ELM for Memory Constrained Edge Computing Devices with Metal Performance Shaders'. Together they form a unique fingerprint.

  • Cite this

    Akusok, A., Leal, L. E., Björk, K-M., & Lendasse, A. (2020). High-Performance ELM for Memory Constrained Edge Computing Devices with Metal Performance Shaders. In J. Cao, C. M. Vong, Y. Miche, & A. Lendasse (Eds.), Proceedings of ELM2019 (pp. 79-88). (Proceedings of ELM2019; Vol. 14). Springer. https://doi.org/10.1007/978-3-030-58989-9_9