Quantum Gradient Class Activation Map for Model Interpretability
This addresses the safety and interpretability problem for quantum machine learning applications, but it is incremental as it adapts classical methods to a quantum setting.
The paper tackles the lack of interpretability in quantum machine learning by proposing a Quantum Gradient Class Activation Map (QGrad-CAM) using Variational Quantum Circuits, resulting in fine-grained, class-discriminative visual explanations for image and speech datasets.
Quantum machine learning (QML) has recently made significant advancements in various topics. Despite the successes, the safety and interpretability of QML applications have not been thoroughly investigated. This work proposes using Variational Quantum Circuits (VQCs) for activation mapping to enhance model transparency, introducing the Quantum Gradient Class Activation Map (QGrad-CAM). This hybrid quantum-classical computing framework leverages both quantum and classical strengths and gives access to the derivation of an explicit formula of feature map importance. Experimental results demonstrate significant, fine-grained, class-discriminative visual explanations generated across both image and speech datasets.