Enabling machine learning across heterogeneous sensor networks with graph autoencoders

Johan Medrano*, Fuchun Joseph Lin

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Machine Learning (ML) has been applied to enable many life-assisting applications, such as abnormality detection in daily routines and automatic emergency request for the solitary elderly. However, in most cases ML algorithms depend on the layout of the target Internet of Things (IoT) sensor network. Hence, to deploy an application across Heterogeneous Sensor Networks (HSNs), i.e. sensor networks with different sensors type or layouts, it is required to repeat the process of data collection and ML algorithm training. In this paper, we introduce a novel framework leveraging deep learning for graphs to enable using the same activity recognition system across HSNs deployed in different smart homes. Using our framework, we were able to transfer activity classifiers trained with activity labels on a source HSN to a target HSN, reaching about 75% of the baseline accuracy on the target HSN without using target activity labels. Moreover, our model can quickly adapt to unseen sensor layouts, which makes it highly suitable for the gradual deployment of real-world ML-based applications. In addition, we show that our framework is resilient to suboptimal graph representations of HSNs.

Original languageEnglish
Title of host publicationAmbient Intelligence - 15th European Conference, AmI 2019, Proceedings
EditorsIoannis Chatzigiannakis, Boris De Ruyter, Irene Mavrommati
PublisherSpringer
Pages153-169
Number of pages17
ISBN (Print)9783030342548
DOIs
StatePublished - 2019
Event15th European Conference on Ambient Intelligence, AmI 2019 - Rome, Italy
Duration: 13 Nov 201915 Nov 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11912 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th European Conference on Ambient Intelligence, AmI 2019
Country/TerritoryItaly
CityRome
Period13/11/1915/11/19

Keywords

  • Graph autoencoders
  • Heterogeneous sensor networks
  • Smart homes

Fingerprint

Dive into the research topics of 'Enabling machine learning across heterogeneous sensor networks with graph autoencoders'. Together they form a unique fingerprint.

Cite this