Practical Insights on Incremental Learning of New Human Physical Activity on the Edge

TL;DR

This paper explores the challenges of incremental learning of human activities on edge devices, emphasizing data storage, computational limits, and class management, based on experiments with the MAGNETO system.

Contribution

It provides practical insights into implementing incremental learning on resource-constrained edge devices for human activity recognition.

Findings

Identified key challenges in edge-based incremental learning.

Demonstrated the impact of storage and computation constraints.

Provided perspectives for future Edge ML development.

Abstract

Edge Machine Learning (Edge ML), which shifts computational intelligence from cloud-based systems to edge devices, is attracting significant interest due to its evident benefits including reduced latency, enhanced data privacy, and decreased connectivity reliance. While these advantages are compelling, they introduce unique challenges absent in traditional cloud-based approaches. In this paper, we delve into the intricacies of Edge-based learning, examining the interdependencies among: (i) constrained data storage on Edge devices, (ii) limited computational power for training, and (iii) the number of learning classes. Through experiments conducted using our MAGNETO system, that focused on learning human activities via data collected from mobile sensors, we highlight these challenges and offer valuable perspectives on Edge ML.