QTabGAN: A Hybrid Quantum-Classical GAN for Tabular Data Synthesis

TL;DR

QTabGAN introduces a hybrid quantum-classical GAN for tabular data synthesis, leveraging an $n$-qubit variational quantum circuit to model complex distributions and a classical mapper to produce tabular samples. The framework supports conditional generation and uses a classical discriminator to drive adversarial training, aiming to preserve real-data distribution and label consistency under privacy constraints. Empirical results across seven real-world datasets show that QTabGAN achieves superior ML utility and statistical similarity metrics compared with state-of-the-art classical baselines and TabularQGAN, highlighting its potential for scalable, quantum-assisted generative modelling on tabular data. The work emphasizes NISQ-compatibility and demonstrates that quantum-enhanced generators can yield high-fidelity synthetic data applicable to privacy-preserving data sharing and downstream learning tasks.

Abstract

Synthesizing realistic tabular data is challenging due to heterogeneous feature types and high dimensionality. We introduce QTabGAN, a hybrid quantum-classical generative adversarial framework for tabular data synthesis. QTabGAN is especially designed for settings where real data are scarce or restricted by privacy constraints. The model exploits the expressive power of quantum circuits to learn complex data distributions, which are then mapped to tabular features using classical neural networks. We evaluate QTabGAN on multiple classification and regression datasets and benchmark it against leading state-of-the-art generative models. Experiments show that QTabGAN achieves up to 54.07% improvement across various classification datasets and evaluation metrics, thus establishing a scalable quantum approach to tabular data synthesis and highlighting its potential for quantum-assisted generative modelling.

Figures (8)

• Figure 1: Basic Architecture of Generative Adversarial Networks (GANs)
• Figure 2: QTabGAN Model Architecture
• Figure 3: Variational Quantum Circuit Design
• Figure 4: Discriminator Architecture
• Figure 5: Evaluation flow for ML Utility of Classification datasets