Home // PATTERNS 2021, The Thirteenth International Conference on Pervasive Patterns and Applications // View article

A New Algorithm Which Runs in Linear Time Enables the Transformation of Legacy Equipment Into Autonomous and Trustworthy IoTs

Authors:
Ole Kristian Ekseth
Erik Morset
Svein-Olaf Hvasshod

Keywords: Approximate Computing, performance, execution- time, signal and image processing, segmentation and clustering, machine learning, algorithms, correlation, similarity-metrics.

Abstract:
The time-cost of today’s classification algorithms are all too high: the use of existing algorithms makes it impossible for cloud-based systems to provide decision-support for remote sensors. Thus, there is a need to develop new algorithms with sufficient accuracy, and with explainable outcomes. Thereby, enabling improved utilization of industrial/physical equipment through smart control. In this work we address this require- ment: this work presents a new methodology for learning, and training, of classification algorithms. The results indicate that the algorithm outperforms existing methods by 10,000x. Importantly, the new algorithm has a memory footprint consid- erably smaller than similar strategies, and is straightforward to validate for trustworthiness. This makes it possible to deploy the algorithm at both IoTs and in the cloud, thereby ensuring its broad applicability.

Pages: 26 to 31

Copyright: Copyright (c) IARIA, 2021

Publication date: April 18, 2021

Published in: conference

ISSN: 2308-3557

ISBN: 978-1-61208-850-1

Location: Porto, Portugal

Dates: from April 18, 2021 to April 22, 2021