Domain adaptation based on hybrid classical-quantum neural network

In the paper, we propose a domain adaptation (DA) scheme based on the hybrid classical-quantum neural network, named hybrid DA. There are four parts, including the classical convolutional neural network part, the label predictor part based on quantum neural network (QNN), the domain classifier part based on QNN, and the gradient reversal layer part. The features from high-dimensional images are extracted by the feature extractor part and then are encoded in the following quantum predictor part and quantum classifier part. After the classical and quantum neural networks are trained jointly, those features that cannot be discriminated between the source and the target domain can be effectively predicted by the quantum predictor. The performance of the proposed DA is verified over DIGIT-5 dataset. The simulation results demonstrate the feasibility of the proposed hybrid DA and show that the hybrid DA has a higher classification accuracy with fewer parameters. It can provide a promising DA in the current era of noisy intermediate-scale quantum devices.

Automated summary

In this paper, a domain adaptation scheme based on hybrid classical-quantum neural network (hybrid DA) is proposed. It consists of four parts: classical convolutional neural network part, label predictor part based on quantum neural network (QNN), domain classifier part based on QNN, and gradient reversal layer part. The proposed hybrid DA can effectively predict features that cannot be discriminated between the source and target domain, and demonstrates higher classification accuracy with fewer parameters compared to classical DA. The feasibility of hybrid DA is verified on DIGIT-5 dataset, showing promising potential in the era of noisy intermediate-scale quantum devices.