LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices

Ioannis Mavromatis; Adrian Sanchez-Mompo; Francesco Raimondo; James Pope; Marcello Bullo; Ingram  Weeks; Vijay Kumar; Pietro Carnelli; George Oikonomou; Theodoros Spyridopoulos; Aftab Khan

doi:10.1109/CCNC51644.2023.10060415

LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices

Ioannis Mavromatis, Adrian Sanchez-Mompo, Francesco Raimondo, James Pope, Marcello Bullo, Ingram Weeks, Vijay Kumar, Pietro Carnelli, George Oikonomou, Theodoros Spyridopoulos, Aftab Khan

Research output: Contribution to conference › Conference Paper › peer-review

11 Citations (Scopus)

Abstract

Data integrity becomes paramount as the number of Internet of Things (IoT) sensor deployments increases. Sensor data can be altered by benign causes or malicious actions. Mechanisms that detect drifts and irregularities can prevent
disruptions and data bias in the state of an IoT application. This paper presents LE3D, an ensemble framework of data drift estimators capable of detecting abnormal sensor behaviours. Working collaboratively with surrounding IoT devices, the type of drift (natural/abnormal) can also be identified and reported
to the end-user. The proposed framework is a lightweight and unsupervised implementation able to run on resource-constrained IoT devices. Our framework is also generalisable, adapting to new sensor streams and environments with minimal online reconfiguration. We compare our method against state-of-the-art
ensemble data drift detection frameworks, evaluating both the real-world detection accuracy as well as the resource utilisation of the implementation. Experimenting with real-world data and emulated drifts, we show the effectiveness of our method, which achieves up to 97% of detection accuracy while requiring minimal resources to run.

Original language	English
Pages	611-619
Number of pages	9
DOIs	https://doi.org/10.1109/CCNC51644.2023.10060415
Publication status	Published - 17 Mar 2023
Event	IEEE Consumer Communications & Networking Conference - Las Vegas, United States Duration: 8 Jan 2023 → 11 Jan 2023 https://ccnc2023.ieee-ccnc.org/

Conference

Conference	IEEE Consumer Communications & Networking Conference
Abbreviated title	CCNC 2023
Country/Territory	United States
City	Las Vegas
Period	8/01/23 → 11/01/23
Internet address	https://ccnc2023.ieee-ccnc.org/

Bibliographical note

Funding Information:
This work was supported in part by Toshiba Europe Ltd. and in part by the SYNERGIA project (grant no. 53707, UK

Publisher Copyright:
© 2023 IEEE.

Access to Document

10.1109/CCNC51644.2023.10060415

Handle.net

Persistent link

11 Citations
1 Conference Paper
1 Article (Academic Journal)

Evaluating Concept Drift Detectors on Real-World Data
Erol, U., Raimondo, F., Pope, J., Gunner, S. D., Kumar, V., Mavromatis, I., Carnelli, P., Spyridopoulos, T., Khan, A. & Oikonomou, G., 29 Jun 2023, (Accepted/In press).
Research output: Contribution to conference › Conference Paper › peer-review
Multi-sensor, Multi-device Smart Building Indoor Environmental Dataset
Erol, U., Raimondo, F., Pope, J., Gunner, S. D., Kumar, V., Mavromatis, I., Carnelli, P., Spyridopoulos, T., Khan, A. & Oikonomou, G., 14 Jul 2023, In: Data in Brief. 49, 109392.
Research output: Contribution to journal › Article (Academic Journal) › peer-review

Open Access
7 Citations (Scopus)

SYNERGIA: 8031 SYNERGIA 53707
Oikonomou, G. (Principal Investigator), Piechocki, R. J. (Co-Investigator), Paquier, T. (Co-Investigator), McConville, R. (Co-Investigator), Pope, J. (Co-Investigator), Raimondo, F. (Researcher), Erol, U. (Researcher), Paschou, C. (Researcher), Zakrzewski, R. (Researcher) & Gunner, S. D. (Researcher)
1/11/20 → 31/10/22
Project: Research

Cite this

Mavromatis, I., Sanchez-Mompo, A., Raimondo, F., Pope, J., Bullo, M., Weeks, I., Kumar, V., Carnelli, P., Oikonomou, G., Spyridopoulos, T., & Khan, A. (2023). LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices. 611-619. Paper presented at IEEE Consumer Communications & Networking Conference, Las Vegas, Nevada, United States. https://doi.org/10.1109/CCNC51644.2023.10060415

@conference{bf5f376ec28244e5a7e4c3f3ae1f8d9f,

title = "LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices",

abstract = "Data integrity becomes paramount as the number of Internet of Things (IoT) sensor deployments increases. Sensor data can be altered by benign causes or malicious actions. Mechanisms that detect drifts and irregularities can prevent disruptions and data bias in the state of an IoT application. This paper presents LE3D, an ensemble framework of data drift estimators capable of detecting abnormal sensor behaviours. Working collaboratively with surrounding IoT devices, the type of drift (natural/abnormal) can also be identified and reported to the end-user. The proposed framework is a lightweight and unsupervised implementation able to run on resource-constrained IoT devices. Our framework is also generalisable, adapting to new sensor streams and environments with minimal online reconfiguration. We compare our method against state-of-the-art ensemble data drift detection frameworks, evaluating both the real-world detection accuracy as well as the resource utilisation of the implementation. Experimenting with real-world data and emulated drifts, we show the effectiveness of our method, which achieves up to 97\% of detection accuracy while requiring minimal resources to run.",

author = "Ioannis Mavromatis and Adrian Sanchez-Mompo and Francesco Raimondo and James Pope and Marcello Bullo and Ingram Weeks and Vijay Kumar and Pietro Carnelli and George Oikonomou and Theodoros Spyridopoulos and Aftab Khan",

note = "Funding Information: This work was supported in part by Toshiba Europe Ltd. and in part by the SYNERGIA project (grant no. 53707, UK Publisher Copyright: {\textcopyright} 2023 IEEE.; IEEE Consumer Communications \& Networking Conference, CCNC 2023 ; Conference date: 08-01-2023 Through 11-01-2023",

year = "2023",

month = mar,

day = "17",

doi = "10.1109/CCNC51644.2023.10060415",

language = "English",

pages = "611--619",

url = "https://ccnc2023.ieee-ccnc.org/",

}

Mavromatis, I, Sanchez-Mompo, A, Raimondo, F , Pope, J, Bullo, M, Weeks, I, Kumar, V, Carnelli, P, Oikonomou, G, Spyridopoulos, T & Khan, A 2023, 'LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices', Paper presented at IEEE Consumer Communications & Networking Conference, Las Vegas, United States, 8/01/23 - 11/01/23 pp. 611-619. https://doi.org/10.1109/CCNC51644.2023.10060415

LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices. / Mavromatis, Ioannis; Sanchez-Mompo, Adrian; Raimondo, Francesco et al.
2023. 611-619 Paper presented at IEEE Consumer Communications & Networking Conference, Las Vegas, Nevada, United States.

Research output: Contribution to conference › Conference Paper › peer-review

TY - CONF

T1 - LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices

AU - Mavromatis, Ioannis

AU - Sanchez-Mompo, Adrian

AU - Raimondo, Francesco

AU - Pope, James

AU - Bullo, Marcello

AU - Weeks, Ingram

AU - Kumar, Vijay

AU - Carnelli, Pietro

AU - Oikonomou, George

AU - Spyridopoulos, Theodoros

AU - Khan, Aftab

PY - 2023/3/17

Y1 - 2023/3/17

N2 - Data integrity becomes paramount as the number of Internet of Things (IoT) sensor deployments increases. Sensor data can be altered by benign causes or malicious actions. Mechanisms that detect drifts and irregularities can prevent disruptions and data bias in the state of an IoT application. This paper presents LE3D, an ensemble framework of data drift estimators capable of detecting abnormal sensor behaviours. Working collaboratively with surrounding IoT devices, the type of drift (natural/abnormal) can also be identified and reported to the end-user. The proposed framework is a lightweight and unsupervised implementation able to run on resource-constrained IoT devices. Our framework is also generalisable, adapting to new sensor streams and environments with minimal online reconfiguration. We compare our method against state-of-the-art ensemble data drift detection frameworks, evaluating both the real-world detection accuracy as well as the resource utilisation of the implementation. Experimenting with real-world data and emulated drifts, we show the effectiveness of our method, which achieves up to 97% of detection accuracy while requiring minimal resources to run.

AB - Data integrity becomes paramount as the number of Internet of Things (IoT) sensor deployments increases. Sensor data can be altered by benign causes or malicious actions. Mechanisms that detect drifts and irregularities can prevent disruptions and data bias in the state of an IoT application. This paper presents LE3D, an ensemble framework of data drift estimators capable of detecting abnormal sensor behaviours. Working collaboratively with surrounding IoT devices, the type of drift (natural/abnormal) can also be identified and reported to the end-user. The proposed framework is a lightweight and unsupervised implementation able to run on resource-constrained IoT devices. Our framework is also generalisable, adapting to new sensor streams and environments with minimal online reconfiguration. We compare our method against state-of-the-art ensemble data drift detection frameworks, evaluating both the real-world detection accuracy as well as the resource utilisation of the implementation. Experimenting with real-world data and emulated drifts, we show the effectiveness of our method, which achieves up to 97% of detection accuracy while requiring minimal resources to run.

U2 - 10.1109/CCNC51644.2023.10060415

DO - 10.1109/CCNC51644.2023.10060415

M3 - Conference Paper

SP - 611

EP - 619

T2 - IEEE Consumer Communications & Networking Conference

Y2 - 8 January 2023 through 11 January 2023

ER -

LE3D: A Lightweight Ensemble Framework of Data Drift Detectors for Resource-Constrained Devices

Abstract

Conference

Bibliographical note

Access to Document

Handle.net

Fingerprint

Research output

Evaluating Concept Drift Detectors on Real-World Data

Multi-sensor, Multi-device Smart Building Indoor Environmental Dataset

Projects

SYNERGIA: 8031 SYNERGIA 53707

Cite this