Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks

Sounak Bhowmik; Travis S. Humble; Himanshu Thapliyal

doi:10.1109/ISVLSI65124.2025.11130199

Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks

Sounak Bhowmik
, Travis S. Humble
, Himanshu Thapliyal

Quantum Science Center

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

Abstract

Quantum neural networks (QNN) hold immense potential for the future of quantum machine learning (QML). However, QNN security and robustness remain largely unexplored. In this work, we proposed novel Trojan attacks based on the quantum computing properties in a QNN-based binary classifier. Our proposed Quantum Properties Trojans (QuPTs) are based on the unitary property of quantum gates to insert noise and Hadamard gates to enable superposition to develop Trojans and attack QNNs. We showed that the proposed QuPTs are significantly stealthier and heavily impact the quantum circuits' performance, specifically QNNs. The most impactful QuPT caused a deterioration of 23% accuracy of the compromised QNN under the experimental setup. To the best of our knowledge, this is the first work on the Trojan attack on a fully quantum neural network independent of any hybrid classical-quantum architecture.

Original language	English
Title of host publication	IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 - Conference Proceedings
Publisher	IEEE Computer Society
ISBN (Electronic)	9798331534776
DOIs	https://doi.org/10.1109/ISVLSI65124.2025.11130199
State	Published - 2025
Event	28th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 - Kalamata, Greece Duration: Jul 6 2025 → Jul 9 2025

Publication series

Name	Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI
ISSN (Print)	2159-3469
ISSN (Electronic)	2159-3477

Conference

Conference	28th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025
Country/Territory	Greece
City	Kalamata
Period	07/6/25 → 07/9/25

Funding

This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-00OR22725.

Keywords

QTrojan
Quantum Neural Network
Quantum Security

Access to Document

10.1109/ISVLSI65124.2025.11130199

Cite this

@inproceedings{4a19912260344e8daa3e43fdd78ecbc8,

title = "Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks",

abstract = "Quantum neural networks (QNN) hold immense potential for the future of quantum machine learning (QML). However, QNN security and robustness remain largely unexplored. In this work, we proposed novel Trojan attacks based on the quantum computing properties in a QNN-based binary classifier. Our proposed Quantum Properties Trojans (QuPTs) are based on the unitary property of quantum gates to insert noise and Hadamard gates to enable superposition to develop Trojans and attack QNNs. We showed that the proposed QuPTs are significantly stealthier and heavily impact the quantum circuits' performance, specifically QNNs. The most impactful QuPT caused a deterioration of 23\% accuracy of the compromised QNN under the experimental setup. To the best of our knowledge, this is the first work on the Trojan attack on a fully quantum neural network independent of any hybrid classical-quantum architecture.",

keywords = "QTrojan, Quantum Neural Network, Quantum Security",

author = "Sounak Bhowmik and Humble, \{Travis S.\} and Himanshu Thapliyal",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 28th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 ; Conference date: 06-07-2025 Through 09-07-2025",

year = "2025",

doi = "10.1109/ISVLSI65124.2025.11130199",

language = "English",

series = "Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI",

publisher = "IEEE Computer Society",

booktitle = "IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 - Conference Proceedings",

}

Bhowmik, S, Humble, TS & Thapliyal, H 2025, Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks. in IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 - Conference Proceedings. Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI, IEEE Computer Society, 28th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025, Kalamata, Greece, 07/6/25. https://doi.org/10.1109/ISVLSI65124.2025.11130199

Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks. / Bhowmik, Sounak; Humble, Travis S.; Thapliyal, Himanshu.
IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 - Conference Proceedings. IEEE Computer Society, 2025. (Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI).

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

TY - GEN

T1 - Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks

AU - Bhowmik, Sounak

AU - Humble, Travis S.

AU - Thapliyal, Himanshu

PY - 2025

Y1 - 2025

N2 - Quantum neural networks (QNN) hold immense potential for the future of quantum machine learning (QML). However, QNN security and robustness remain largely unexplored. In this work, we proposed novel Trojan attacks based on the quantum computing properties in a QNN-based binary classifier. Our proposed Quantum Properties Trojans (QuPTs) are based on the unitary property of quantum gates to insert noise and Hadamard gates to enable superposition to develop Trojans and attack QNNs. We showed that the proposed QuPTs are significantly stealthier and heavily impact the quantum circuits' performance, specifically QNNs. The most impactful QuPT caused a deterioration of 23% accuracy of the compromised QNN under the experimental setup. To the best of our knowledge, this is the first work on the Trojan attack on a fully quantum neural network independent of any hybrid classical-quantum architecture.

AB - Quantum neural networks (QNN) hold immense potential for the future of quantum machine learning (QML). However, QNN security and robustness remain largely unexplored. In this work, we proposed novel Trojan attacks based on the quantum computing properties in a QNN-based binary classifier. Our proposed Quantum Properties Trojans (QuPTs) are based on the unitary property of quantum gates to insert noise and Hadamard gates to enable superposition to develop Trojans and attack QNNs. We showed that the proposed QuPTs are significantly stealthier and heavily impact the quantum circuits' performance, specifically QNNs. The most impactful QuPT caused a deterioration of 23% accuracy of the compromised QNN under the experimental setup. To the best of our knowledge, this is the first work on the Trojan attack on a fully quantum neural network independent of any hybrid classical-quantum architecture.

KW - QTrojan

KW - Quantum Neural Network

KW - Quantum Security

UR - https://www.scopus.com/pages/publications/105016177933

U2 - 10.1109/ISVLSI65124.2025.11130199

DO - 10.1109/ISVLSI65124.2025.11130199

M3 - Conference contribution

AN - SCOPUS:105016177933

T3 - Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI

BT - IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025 - Conference Proceedings

PB - IEEE Computer Society

T2 - 28th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2025

Y2 - 6 July 2025 through 9 July 2025

ER -

Quantum Properties Trojans (QuPTs) for Attacking Quantum Neural Networks

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this