Deep Spiking Quantum Neural Network for Noisy Image Classification

stochastic-based modelling, promising that the inherent uncertainty in quantum computing will prove to be a significant advantage, driving quantum and neuromorphic computing research
to new heights. Spiking Neural Networks (SNNs) are gaining popularity due to their inherent ability to process spatial and temporal data. However, it is a daunting task to train
the network weights of classical SNN due to the stochastic behaviour of neuron signals and the inherent non-differentiable spike events. This paper introduces a supervised Deep Spiking
Quantum Neural Network (DSQ-Net) using a hybrid classicalquantum framework having the merits of amplitude encoding in a dressed quantum layer. A novel attempt has been made
to obviate the challenges in training a classical SNN, assisted by a Variational Quantum Circuit (VQC) in the proposed hybrid classical-quantum framework. The proposed DSQ-Net
is rigorously validated and benchmarked on the ideal PennyLane Quantum Simulator with limited quantum hardware. The experiments have been conducted on unseen test images with imposed noise from the FashionMNIST, MNIST, KMNIST and CIFAR-10 datasets. Classification accuracy is reported to be 95.6% for the proposed DSQ-Net model and
it outperforms the classical counterpart (Deep Spiking Neural Networks), shallow Random Quantum Neural Networks (RQNN), ResNet-18 and AlexNet. The PyTorch implementation
of DSQ-Net is made available on Github0:https://anonymous.4open.science/r/DSQ-Net-037E.