In recent years we are witnessing a rapid increase in the diffusion of the Internet of things (IoT) technology, with a large scale adoption of interconnected heterogeneous devices that are pervasively collecting information through the interaction with humans in their environment. The adoption of Machine Learning (ML) methodologies can play a fundamental role, allowing smarter IoT applications to continuously adapt to evolving environmental conditions and user’s needs. In this context, the time is now ripe for a decisive step forward in the direction of a systematic integration of ML functionalities within the IoT platform. In this paper, we outline the principles that should guide the realization of a ML service for the IoT, proposing a conceptual architecture of such a learning service, integrated within the IoT reference model. Our proposal leverages on the experience of recent successful European initiatives that led to the realization of intelligent sensor networks built on the synergy between resource efficient ML models for temporal data processing and wireless sensor networks. The relevant impact of ML in applicative domains of interest for the IoT is also enucleated through a brief summary of recent results.

On the Need of Machine Learning as a Service for the Internet of Things

Davide Bacciu;Stefano Chessa;Claudio Gallicchio;Alessio Micheli
2017

Abstract

In recent years we are witnessing a rapid increase in the diffusion of the Internet of things (IoT) technology, with a large scale adoption of interconnected heterogeneous devices that are pervasively collecting information through the interaction with humans in their environment. The adoption of Machine Learning (ML) methodologies can play a fundamental role, allowing smarter IoT applications to continuously adapt to evolving environmental conditions and user’s needs. In this context, the time is now ripe for a decisive step forward in the direction of a systematic integration of ML functionalities within the IoT platform. In this paper, we outline the principles that should guide the realization of a ML service for the IoT, proposing a conceptual architecture of such a learning service, integrated within the IoT reference model. Our proposal leverages on the experience of recent successful European initiatives that led to the realization of intelligent sensor networks built on the synergy between resource efficient ML models for temporal data processing and wireless sensor networks. The relevant impact of ML in applicative domains of interest for the IoT is also enucleated through a brief summary of recent results.
978-1-4503-5243-7
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11568/890677
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 11
  • ???jsp.display-item.citation.isi??? 4
social impact