Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes

Yang Wai Chow; Willy Susilo; Jianfeng Wang; Richard Buckland; Joonsang Baek; Jongkil Kim; Nan Li

doi:10.1007/978-3-030-30619-9_23

Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes

Yang Wai Chow, Willy Susilo, Jianfeng Wang, Richard Buckland, Joonsang Baek, Jongkil Kim, Nan Li

Department of Cyber Security

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

1 Scopus citations

Abstract

Machine learning is becoming increasingly popular in a variety of modern technology. However, research has demonstrated that machine learning models are vulnerable to adversarial examples in their inputs. Potential attacks include poisoning datasets by perturbing input samples to mislead a machine learning model into producing undesirable results. Such perturbations are often subtle and imperceptible from a human’s perspective. This paper investigates two methods of verifying the visual fidelity of image based datasets by detecting perturbations made to the data using QR codes. In the first method, a verification string is stored for each image in a dataset. These verification strings can be used to determine whether an image in the dataset has been perturbed. In the second method, only a single verification string stored and is used to verify whether an entire dataset is intact.

Original language	English
Title of host publication	Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings
Editors	Xiaofeng Chen, Xinyi Huang, Jun Zhang
Publisher	Springer Verlag
Pages	320-335
Number of pages	16
ISBN (Print)	9783030306182
DOIs	https://doi.org/10.1007/978-3-030-30619-9_23
State	Published - 2019
Event	2nd International Conference on Machine Learning for Cyber Security, ML4CS 2019 - Xi'an, China Duration: 19 Sep 2019 → 21 Sep 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11806 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	2nd International Conference on Machine Learning for Cyber Security, ML4CS 2019
Country/Territory	China
City	Xi'an
Period	19/09/19 → 21/09/19

Bibliographical note

Publisher Copyright:
© 2019, Springer Nature Switzerland AG.

Keywords

Adversarial machine learning
Cyber security
QR code
Visual fidelity
Watermarking

Access to Document

10.1007/978-3-030-30619-9_23

Cite this

Chow, Y. W., Susilo, W., Wang, J., Buckland, R., Baek, J., Kim, J., & Li, N. (2019). Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes. In X. Chen, X. Huang, & J. Zhang (Eds.), Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings (pp. 320-335). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11806 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-30619-9_23

Chow, Yang Wai ; Susilo, Willy ; Wang, Jianfeng et al. / Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes. Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings. editor / Xiaofeng Chen ; Xinyi Huang ; Jun Zhang. Springer Verlag, 2019. pp. 320-335 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{a87e1735f1d6401ebbcea27249444573,

title = "Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes",

abstract = "Machine learning is becoming increasingly popular in a variety of modern technology. However, research has demonstrated that machine learning models are vulnerable to adversarial examples in their inputs. Potential attacks include poisoning datasets by perturbing input samples to mislead a machine learning model into producing undesirable results. Such perturbations are often subtle and imperceptible from a human{\textquoteright}s perspective. This paper investigates two methods of verifying the visual fidelity of image based datasets by detecting perturbations made to the data using QR codes. In the first method, a verification string is stored for each image in a dataset. These verification strings can be used to determine whether an image in the dataset has been perturbed. In the second method, only a single verification string stored and is used to verify whether an entire dataset is intact.",

keywords = "Adversarial machine learning, Cyber security, QR code, Visual fidelity, Watermarking",

author = "Chow, {Yang Wai} and Willy Susilo and Jianfeng Wang and Richard Buckland and Joonsang Baek and Jongkil Kim and Nan Li",

note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 2nd International Conference on Machine Learning for Cyber Security, ML4CS 2019 ; Conference date: 19-09-2019 Through 21-09-2019",

year = "2019",

doi = "10.1007/978-3-030-30619-9_23",

language = "English",

isbn = "9783030306182",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "320--335",

editor = "Xiaofeng Chen and Xinyi Huang and Jun Zhang",

booktitle = "Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings",

}

Chow, YW, Susilo, W, Wang, J, Buckland, R, Baek, J, Kim, J & Li, N 2019, Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes. in X Chen, X Huang & J Zhang (eds), Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11806 LNCS, Springer Verlag, pp. 320-335, 2nd International Conference on Machine Learning for Cyber Security, ML4CS 2019, Xi'an, China, 19/09/19. https://doi.org/10.1007/978-3-030-30619-9_23

Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes. / Chow, Yang Wai; Susilo, Willy; Wang, Jianfeng et al.
Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings. ed. / Xiaofeng Chen; Xinyi Huang; Jun Zhang. Springer Verlag, 2019. p. 320-335 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11806 LNCS).

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

TY - GEN

T1 - Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes

AU - Chow, Yang Wai

AU - Susilo, Willy

AU - Wang, Jianfeng

AU - Buckland, Richard

AU - Baek, Joonsang

AU - Kim, Jongkil

AU - Li, Nan

PY - 2019

Y1 - 2019

N2 - Machine learning is becoming increasingly popular in a variety of modern technology. However, research has demonstrated that machine learning models are vulnerable to adversarial examples in their inputs. Potential attacks include poisoning datasets by perturbing input samples to mislead a machine learning model into producing undesirable results. Such perturbations are often subtle and imperceptible from a human’s perspective. This paper investigates two methods of verifying the visual fidelity of image based datasets by detecting perturbations made to the data using QR codes. In the first method, a verification string is stored for each image in a dataset. These verification strings can be used to determine whether an image in the dataset has been perturbed. In the second method, only a single verification string stored and is used to verify whether an entire dataset is intact.

AB - Machine learning is becoming increasingly popular in a variety of modern technology. However, research has demonstrated that machine learning models are vulnerable to adversarial examples in their inputs. Potential attacks include poisoning datasets by perturbing input samples to mislead a machine learning model into producing undesirable results. Such perturbations are often subtle and imperceptible from a human’s perspective. This paper investigates two methods of verifying the visual fidelity of image based datasets by detecting perturbations made to the data using QR codes. In the first method, a verification string is stored for each image in a dataset. These verification strings can be used to determine whether an image in the dataset has been perturbed. In the second method, only a single verification string stored and is used to verify whether an entire dataset is intact.

KW - Adversarial machine learning

KW - Cyber security

KW - QR code

KW - Visual fidelity

KW - Watermarking

UR - http://www.scopus.com/inward/record.url?scp=85072862952&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-30619-9_23

DO - 10.1007/978-3-030-30619-9_23

M3 - Conference contribution

AN - SCOPUS:85072862952

SN - 9783030306182

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 320

EP - 335

BT - Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings

A2 - Chen, Xiaofeng

A2 - Huang, Xinyi

A2 - Zhang, Jun

PB - Springer Verlag

T2 - 2nd International Conference on Machine Learning for Cyber Security, ML4CS 2019

Y2 - 19 September 2019 through 21 September 2019

ER -

Chow YW, Susilo W, Wang J, Buckland R, Baek J, Kim J et al. Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes. In Chen X, Huang X, Zhang J, editors, Machine Learning for Cyber Security - 2nd International Conference, ML4CS 2019, Proceedings. Springer Verlag. 2019. p. 320-335. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-30619-9_23

Protecting the Visual Fidelity of Machine Learning Datasets Using QR Codes

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this