Quantum GAN for Bars-and-Stripes

Q2 2025–26 · Quantum Computation and Quantum Programming (University of Leiden)

Three-student group project in which we implemented a Quantum Generative Adversarial Network (QGAN) for generating Bars-and-Stripes patterns. The generator was a Parametrized Quantum Circuit (PQC) adversarially trained by a discriminator realized as a classical Neural Network. After training and performance benchmarking we also explored topics such as choosing the right ansatz for the PQC and using the Wasserstein distance in the loss function.