CopyQNN: Quantum Neural Network Extraction Attack under Varying Quantum Noise
Zhenxiao Fu * Leyi Zhao * Xuhong Zhang * Yilun Xu † Gang Huang † Fan Chen * * Indiana University Bloomington † Lawrence Berkeley National Laboratory * {zhfu, leyizhao, zhangxuh, fc7}@iu.edu † {yilunxu, ghuang}@lbl.gov · 2025
Read CopyQNN: Quantum Neural Network Extraction Attack under Varying Quantum Noise by Zhenxiao Fu * Leyi Zhao * Xuhong Zhang * Yilun Xu † Gang Huang † Fan Chen * * Indiana University Bloomington † Lawrence Berkeley National Laboratory * {zhfu, leyizhao, zhangxuh, fc7}@iu.edu † {yilunxu, ghuang}@lbl.gov (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-00366.txt.
Description
Zhenxiao Fu and co-authors study an attack that copies quantum neural networks from cloud services, in a 2025 paper. Trained quantum models are valuable property and are often rented out online. Attackers can query a service, collect labeled answers, and train a local copy. Past attack work ignored a key fact of real quantum hardware: noise levels vary.
Their framework, CopyQNN, cleans the collected data in three steps based on noise sensitivity. It then joins two learning tricks inside the quantum model to sharpen the copy. The aim is a stand-in model that mimics the cloud model even under shifting noise. The work shows a risk for quantum cloud services and hints at where defenses are needed.
Catalog record
| CREATOR | Zhenxiao Fu * Leyi Zhao * Xuhong Zhang * Yilun Xu † Gang Huang † Fan Chen * * Indiana University Bloomington † Lawrence Berkeley National Laboratory * {zhfu, leyizhao, zhangxuh, fc7}@iu.edu † {yilunxu, ghuang}@lbl.gov |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-00366