Saved in:
Bibliographic Details
Main Authors: Ordóñez, Sebastián A. Cajas, Samanta, Jaydeep, Suárez-Cetrulo, Andrés L., Carbajo, Ricardo Simón
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.18634
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912290518859776
author Ordóñez, Sebastián A. Cajas
Samanta, Jaydeep
Suárez-Cetrulo, Andrés L.
Carbajo, Ricardo Simón
author_facet Ordóñez, Sebastián A. Cajas
Samanta, Jaydeep
Suárez-Cetrulo, Andrés L.
Carbajo, Ricardo Simón
contents The Internet of Things is an example domain where data is perpetually generated in ever-increasing quantities, reflecting the proliferation of connected devices and the formation of continuous data streams over time. Consequently, the demand for ad-hoc, cost-effective machine learning solutions must adapt to this evolving data influx. This study tackles the task of offloading in small gateways, exacerbated by their dynamic availability over time. An approach leveraging CPU utilization metrics using online and continual machine learning techniques is proposed to predict gateway availability. These methods are compared to popular machine learning algorithms and a recent time-series foundation model, Lag-Llama, for fine-tuned and zero-shot setups. Their performance is benchmarked on a dataset of CPU utilization measurements over time from an IoT gateway and focuses on model metrics such as prediction errors, training and inference times, and memory consumption. Our primary objective is to study new efficient ways to predict CPU performance in IoT environments. Across various scenarios, our findings highlight that ensemble and online methods offer promising results for this task in terms of accuracy while maintaining a low resource footprint.
format Preprint
id arxiv_https___arxiv_org_abs_2503_18634
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Adaptive Machine Learning for Resource-Constrained Environments
Ordóñez, Sebastián A. Cajas
Samanta, Jaydeep
Suárez-Cetrulo, Andrés L.
Carbajo, Ricardo Simón
Machine Learning
68T05, 62M20
I.2.6; H.2.8
The Internet of Things is an example domain where data is perpetually generated in ever-increasing quantities, reflecting the proliferation of connected devices and the formation of continuous data streams over time. Consequently, the demand for ad-hoc, cost-effective machine learning solutions must adapt to this evolving data influx. This study tackles the task of offloading in small gateways, exacerbated by their dynamic availability over time. An approach leveraging CPU utilization metrics using online and continual machine learning techniques is proposed to predict gateway availability. These methods are compared to popular machine learning algorithms and a recent time-series foundation model, Lag-Llama, for fine-tuned and zero-shot setups. Their performance is benchmarked on a dataset of CPU utilization measurements over time from an IoT gateway and focuses on model metrics such as prediction errors, training and inference times, and memory consumption. Our primary objective is to study new efficient ways to predict CPU performance in IoT environments. Across various scenarios, our findings highlight that ensemble and online methods offer promising results for this task in terms of accuracy while maintaining a low resource footprint.
title Adaptive Machine Learning for Resource-Constrained Environments
topic Machine Learning
68T05, 62M20
I.2.6; H.2.8
url https://arxiv.org/abs/2503.18634