Abstract:As an emerging field that aims to bridge the gap between human activities and computing systems, human-centered computing (HCC) in cloud, edge, fog has had a huge impact on the artificial intelligence algorithms. The quantum generative adversarial network (QGAN) is considered to be one of the quantum machine learning algorithms with great application prospects, which also should be improved to conform to the human-centered paradigm. The generation process of QGAN is relatively random and the generated model does not conform to the human-centered concept, so it is not quite suitable for real scenarios. In order to solve these problems, a hybrid quantum-classical conditional generative adversarial network (QCGAN) algorithm is proposed, which is a knowledge-driven human-computer interaction computing mode that can be implemented in cloud. The purposes of stabilizing the generation process and realizing the interaction between human and computing process are achieved by inputting artificial conditional information in the generator and discriminator. The generator uses the parameterized quantum circuit with an all-to-all connected topology, which facilitates the tuning of network parameters during the training process. The discriminator uses the classical neural network, which effectively avoids the "input bottleneck" of quantum machine learning. Finally, the BAS training set is selected to conduct experiment on the quantum cloud computing platform. The result shows that the QCGAN algorithm can effectively converge to the Nash equilibrium point after training and perform human-centered classification generation tasks.

What problem does this paper attempt to address?

The main problems that this paper attempts to solve are the randomness and uncontrollability in the generation process of quantum generative adversarial networks (QGAN), and the generated models do not conform to the human - centered concept, making it difficult to be applied to practical scenarios. To solve these problems, the paper proposes a hybrid quantum - classical algorithm (QCGAN) based on conditional generative adversarial networks (CGAN), aiming to stabilize the generation process by introducing artificial conditional information and realize the interaction between humans and the computing process. Specifically: - **Problem Background**: Traditional QGAN has high randomness when generating data, and the generated data does not fully meet the conditions or goals set by humans, which limits its applicability in practical applications, especially in fields related to human activities that require high - level customization and control. - **Solution**: By adding conditional constraints (i.e., conditional variables) on the basis of QGAN to guide the learning process of the generator and the discriminator, the generated data can better meet specific conditions, thereby improving the application value and practicality of the algorithm. - **Technical Means**: Use parameterized quantum circuits (PQC) as generators, and use the fully - connected topology to easily adjust network parameters; the discriminator adopts a classical neural network, which effectively avoids the "input bottleneck" problem in quantum machine learning. In addition, by simultaneously inputting artificial conditional information into the generator and the discriminator, the interaction between humans and the computing process is realized, and the controllability of the generation process and the quality of the generated data are improved. The paper verifies the effectiveness of the QCGAN algorithm through experiments, indicating that the algorithm can effectively converge to the Nash equilibrium point after training and successfully complete human - centered classification generation tasks. This not only proves the theoretical feasibility of QCGAN, but also provides new ideas and technical support for the application of quantum computing in the field of generative adversarial networks.

A hybrid quantum-classical conditional generative adversarial network algorithm for human-centered paradigm in cloud

LatentQGAN: A Hybrid QGAN with Classical Convolutional Autoencoder

Variational Quantum Circuits Enhanced Generative Adversarial Network

A Hybrid Quantum-Classical Generative Adversarial Network for Near-Term Quantum Processors

Quantum Generative Adversarial Networks in a Silicon Photonic Chip with Maximum Expressibility

Quantum Deep Generative Prior with Programmable Quantum Circuits

Hybrid Quantum-Classical Normalizing Flow

Detection and evaluation of abnormal user behavior based on quantum generation adversarial network

Experimental Quantum Generative Adversarial Networks for Image Generation

Quantum generative adversarial network for generating discrete distribution

Quantum Generative Adversarial Networks: Bridging Classical and Quantum Realms

Entangling Quantum Generative Adversarial Networks

Quantum generative adversarial networks with multiple superconducting qubits

Re-QGAN: an optimized adversarial quantum circuit learning framework

Learning and Inference on Generative Adversarial Quantum Circuits

Quantum Generative Adversarial Networks: Generating and Detecting Quantum Product States

Quantum Generative Models for Data Generation

Generative AI-enabled Quantum Computing Networks and Intelligent Resource Allocation

iHQGAN: A Lightweight Invertible Hybrid Quantum-Classical Generative Adversarial Network for Unsupervised Image-to-Image Translation

Unrolled generative adversarial network for continuous distributions under hybrid quantum-classical model