On the Generalization of Adversarially Trained Quantum Classifiers
This work addresses the robustness of quantum machine learning models against adversarial threats, providing theoretical insights for researchers in quantum computing and security, though it is incremental as it builds on existing adversarial training concepts.
The paper tackles the vulnerability of quantum classifiers to adversarial attacks by establishing generalization error bounds for adversarially trained quantum classifiers, showing that the excess error scales as 1/√m with training sample size and depends on input dimension and embedding choices.
Quantum classifiers are vulnerable to adversarial attacks that manipulate their input classical or quantum data. A promising countermeasure is adversarial training, where quantum classifiers are trained by using an attack-aware, adversarial loss function. This work establishes novel bounds on the generalization error of adversarially trained quantum classifiers when tested in the presence of perturbation-constrained adversaries. The bounds quantify the excess generalization error incurred to ensure robustness to adversarial attacks as scaling with the training sample size $m$ as $1/\sqrt{m}$, while yielding insights into the impact of the quantum embedding. For quantum binary classifiers employing \textit{rotation embedding}, we find that, in the presence of adversarial attacks on classical inputs $\mathbf{x}$, the increase in sample complexity due to adversarial training over conventional training vanishes in the limit of high dimensional inputs $\mathbf{x}$. In contrast, when the adversary can directly attack the quantum state $ρ(\mathbf{x})$ encoding the input $\mathbf{x}$, the excess generalization error depends on the choice of embedding only through its Hilbert space dimension. The results are also extended to multi-class classifiers. We validate our theoretical findings with numerical experiments.