QML-Essentials

Quantum Fourier models provide an intuitive an mathematically solid basis for exploring capabilities and limitations of quantum neural networks. Representing these kind of models via a truncated Fourier series (hence the name), allows applying all well-known methods from classical signal processing.

To facilitate an easy entry into this research area and to provide an open-source implementation of many known metrics and circuits, we actively develop the QML-Essentials at SCC, KIT in collaboration with the Laboratory for Digitalization at OTH Regensburg.

Get started with

uv add qml-essentials

and explore the capabilities of quantum Fourier models (the focus of my PHD btw.). We provide some tutorials on our website that help you setting up a training pipeline with these models in no time.

In our paper, we introduce the framework and presented our initial set of features, covering

• calculation of entanglement using the Meyer-Wallach method and Bell measurements
• calculation of expressibility using the KL-divergence to Haar random states
• frequency analysis of a Fourier model using both analytical and numerical method
• a set predefined circuits, commonly used in the literature
• .. and much more

Feel free to check it out in the proceedings to QSW25 where we got this work accepted.