Physical Review Research · 2023 · 60 citations · 25 references
ComputingComputer ScienceArtificial IntelligenceMachine LearningQuantum ComputingQuantum AlgorithmsQuantum ComputationQuantum Machine LearningAdversarial Attack DetectionAdversarial RobustnessQuantum AdvantageQuantum BenchmarkingQuantum Machine Learning AlgorithmsQuantum NetworksQuantum Neural NetworksRobustness
Machine learning methods such as artificial neural networks are ubiquitous, yet they can be easily fooled by adversarial attacks, and the extent of such vulnerabilities in quantum machine learning remains unclear. This study benchmarks the robustness of quantum variational classifiers at scale and proposes a novel adversarial attack detection technology that combines quantum and classical network outcomes. The authors train QVCs on both simple and complex image datasets, subject them to a range of high‑end adversarial attacks, and integrate the quantum and classical predictions to develop the detection method. Results show that QVCs exhibit markedly greater robustness against classical attacks by learning features invisible to classical nets, while attacks on quantum networks can also deceive classical nets, revealing a new quantum advantage through superior robustness that could enhance security in autonomous vehicles, cybersecurity, and surveillance systems.
Machine learning (ML) methods such as artificial neural networks are rapidly becoming ubiquitous in modern science, technology and industry. Despite their accuracy and sophistication, neural networks can be easily fooled by carefully designed malicious inputs known as adversarial attacks. While such vulnerabilities remain a serious challenge for classical neural networks, the extent of their existence is not fully understood in the quantum ML setting. In this work, we benchmark the robustness of quantum ML networks, such as quantum variational classifiers (QVC), at scale by performing rigorous training for both simple and complex image datasets and through a variety of high-end adversarial attacks. Our results show that QVCs offer a notably enhanced robustness against classical adversarial attacks by learning features which are not detected by the classical neural networks, indicating a possible quantum advantage for ML tasks. Contrarily, and remarkably, the converse is not true, with attacks on quantum networks also capable of deceiving classical neural networks. By combining quantum and classical network outcomes, we propose a novel adversarial attack detection technology. Traditionally quantum advantage in ML systems has been sought through increased accuracy or algorithmic speed-up, but our work has revealed the potential for a new kind of quantum advantage through superior robustness of ML models, whose practical realisation will address serious security concerns and reliability issues of ML algorithms employed in a myriad of applications including autonomous vehicles, cybersecurity, and surveillance robotic systems.
25
Gradient-based learning applied to document recognition
Yann LeCun, Léon Bottou, Yoshua Bengio et al. · Proceedings of the IEEE · 1998 · 58.2K citations · Full text
Highly accurate protein structure prediction with AlphaFold
John Jumper, Alexander Pritzel, Tim Green et al. · Nature · 2021 · 45.2K citations · Full text
Advances in neural information processing systems 7
Computers & Mathematics with Applications · 1996 · 14.4K citations
Proceedings of IEEE International Conference on Computer Vision
1995 · 12.9K citations
Jacob Biamonte, Péter Wittek, Nicola Pancotti et al. · Nature · 2017 · 4.4K citations · Full text