Camouflage Learning

Stephan Sigg; Le Ngu Nguyen; Jing Ma

doi:10.1109/PerComWorkshops51409.2021.9431111

Camouflage Learning

Stephan Sigg, Le Ngu Nguyen, Jing Ma

Ambient Intelligence

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

42 Downloads (Pure)

Abstract

Federated learning has been proposed as a concept for distributed machine learning which enforces privacy by avoiding sharing private data with a coordinator or distributed nodes. However, information on local data might be leaked through the model updates. We propose Camouflage learning, a machine learning scheme that distributes both the data and the model. Neither the distributed devices nor the coordinator is at any point in time in possession of the complete model. Furthermore, data and model are obfuscated during distributed model inference and distributed model training. Camouflage learning can be implemented with various Machine learning schemes.

Original language	English
Title of host publication	2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2021
Publisher	IEEE
Pages	724-729
Number of pages	6
ISBN (Electronic)	978-1-6654-0424-2
DOIs	https://doi.org/10.1109/PerComWorkshops51409.2021.9431111
Publication status	Published - 25 May 2021
MoE publication type	A4 Article in a conference publication
Event	IEEE International Conference on Pervasive Computing and Communications Workshops - Kassel, Germany Duration: 22 Mar 2021 → 26 Mar 2021 https://www.percom.org/

Publication series

Name	IEEE international conference on pervasive computing and communications workshops
Publisher	IEEE

Conference

Conference	IEEE International Conference on Pervasive Computing and Communications Workshops
Abbreviated title	PerCom Workshops
Country/Territory	Germany
City	Kassel
Period	22/03/2021 → 26/03/2021
Internet address	https://www.percom.org/

Keywords

Distributed machine learning
Internet of Things
Multi-key homomorphic encryption
Privacy

Access to Document

10.1109/PerComWorkshops51409.2021.9431111

ELEC_Sigg_etal_Camouflage_Learning_IEEE_PerCom_2021Accepted author manuscript, 2.21 MB

OpenUrl availability

Full text

Cite this

Sigg, S., Nguyen, L. N., & Ma, J. (2021). Camouflage Learning. In 2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2021 (pp. 724-729). [9431111] (IEEE international conference on pervasive computing and communications workshops). IEEE. https://doi.org/10.1109/PerComWorkshops51409.2021.9431111

@inproceedings{6e4b3f904c394176a2857e7fb57b8dbd,

title = "Camouflage Learning",

abstract = "Federated learning has been proposed as a concept for distributed machine learning which enforces privacy by avoiding sharing private data with a coordinator or distributed nodes. However, information on local data might be leaked through the model updates. We propose Camouflage learning, a machine learning scheme that distributes both the data and the model. Neither the distributed devices nor the coordinator is at any point in time in possession of the complete model. Furthermore, data and model are obfuscated during distributed model inference and distributed model training. Camouflage learning can be implemented with various Machine learning schemes. ",

keywords = "Distributed machine learning, Internet of Things, Multi-key homomorphic encryption, Privacy",

author = "Stephan Sigg and Nguyen, {Le Ngu} and Jing Ma",

note = "Funding Information: ACKNOWLEDGMENT We would like to acknowledge partial funding by the Academy of Finland Project ABACUS (ICT 2023). Publisher Copyright: {\textcopyright} 2021 IEEE.; IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops ; Conference date: 22-03-2021 Through 26-03-2021",

year = "2021",

month = may,

day = "25",

doi = "10.1109/PerComWorkshops51409.2021.9431111",

language = "English",

series = "IEEE international conference on pervasive computing and communications workshops",

publisher = "IEEE",

pages = "724--729",

booktitle = "2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2021",

address = "United States",

url = "https://www.percom.org/",

}

Sigg, S, Nguyen, LN & Ma, J 2021, Camouflage Learning. in 2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2021., 9431111, IEEE international conference on pervasive computing and communications workshops, IEEE, pp. 724-729, IEEE International Conference on Pervasive Computing and Communications Workshops, Kassel, Hesse, Germany, 22/03/2021. https://doi.org/10.1109/PerComWorkshops51409.2021.9431111

Camouflage Learning. / Sigg, Stephan; Nguyen, Le Ngu; Ma, Jing.
2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2021. IEEE, 2021. p. 724-729 9431111 (IEEE international conference on pervasive computing and communications workshops).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Camouflage Learning

AU - Sigg, Stephan

AU - Nguyen, Le Ngu

AU - Ma, Jing

PY - 2021/5/25

Y1 - 2021/5/25

N2 - Federated learning has been proposed as a concept for distributed machine learning which enforces privacy by avoiding sharing private data with a coordinator or distributed nodes. However, information on local data might be leaked through the model updates. We propose Camouflage learning, a machine learning scheme that distributes both the data and the model. Neither the distributed devices nor the coordinator is at any point in time in possession of the complete model. Furthermore, data and model are obfuscated during distributed model inference and distributed model training. Camouflage learning can be implemented with various Machine learning schemes.

AB - Federated learning has been proposed as a concept for distributed machine learning which enforces privacy by avoiding sharing private data with a coordinator or distributed nodes. However, information on local data might be leaked through the model updates. We propose Camouflage learning, a machine learning scheme that distributes both the data and the model. Neither the distributed devices nor the coordinator is at any point in time in possession of the complete model. Furthermore, data and model are obfuscated during distributed model inference and distributed model training. Camouflage learning can be implemented with various Machine learning schemes.

KW - Distributed machine learning

KW - Internet of Things

KW - Multi-key homomorphic encryption

KW - Privacy

UR - http://www.scopus.com/inward/record.url?scp=85107590123&partnerID=8YFLogxK

U2 - 10.1109/PerComWorkshops51409.2021.9431111

DO - 10.1109/PerComWorkshops51409.2021.9431111

M3 - Conference contribution

AN - SCOPUS:85107590123

T3 - IEEE international conference on pervasive computing and communications workshops

SP - 724

EP - 729

BT - 2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2021

PB - IEEE

T2 - IEEE International Conference on Pervasive Computing and Communications Workshops

Y2 - 22 March 2021 through 26 March 2021

ER -

Camouflage Learning

Abstract

Publication series

Conference

Keywords

Access to Document

OpenUrl availability

Other files and links

Fingerprint

Adaptive ambient backscatter Communications for ultra-low power Systems

Cite this

Camouflage Learning

Abstract

Publication series

Conference

Keywords

Access to Document

OpenUrl availability

Other files and links

Fingerprint

Projects

Adaptive ambient backscatter Communications for ultra-low power Systems

Cite this