Adversarial Robustness of Partitioned Quantum Classifiers
This addresses adversarial robustness for quantum classifiers in the NISQ era, which is incremental as it builds on existing classical adversarial attack methods applied to a specific quantum technique.
The paper investigates how partitioning quantum classifiers via circuit cutting increases their vulnerability to adversarial attacks, linking attacks on state preparation channels to adversarial gates in intermediate layers, and finds that this susceptibility grows with the number of partitions, with experimental results showing a 20% increase in attack success rate for a 4-partition classifier.
Adversarial robustness in quantum classifiers is a critical area of study, providing insights into their performance compared to classical models and uncovering potential advantages inherent to quantum machine learning. In the NISQ era of quantum computing, circuit cutting is a notable technique for simulating circuits that exceed the qubit limitations of current devices, enabling the distribution of a quantum circuit's execution across multiple quantum processing units through classical communication. We examine how partitioning quantum classifiers through circuit cutting increase their susceptibility to adversarial attacks, establishing a link between attacking the state preparation channels in wire cutting and implementing adversarial gates within intermediate layers of a quantum classifier. We then proceed to study the latter problem from both a theoretical and experimental perspective.