SAMOP 2023 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

QI: Fachverband Quanteninformation

QI 3: Quantum Machine Learning

QI 3.5: Vortrag

Montag, 6. März 2023, 12:15–12:30, B305

Quantum kernel methods for regression — •Jan Schnabel — Fraunhofer-Institut für Produktionstechnik und Automatisierung IPA, Center for Cyber Cognitive Intelligence (CCI), Stuttgart, Germany

It was shown in Refs. [1,2] that encoding data into a quantum state and interpreting the respective expectation value when measuring w.r.t. an observable as machine learning model, links quantum computing to the rich framework of classical kernel theory. Hence, these theoretical tools can now be used to understand quantum models. Here, the inherent structure of quantum kernel methods is particularly suited for NISQ applications. As a result, these facts caused constantly growing research activities in this field, where little attention has been hitherto paid to quantum kernel regression problems.

In this talk, I briefly introduce the core theoretical concepts of different approaches for computing quantum kernels before discussing associated challenges. The latter includes the role of classical data pre-processing and selection, data redundancies as well as the design of quantum feature maps. These aspects are discussed based on project-specific use cases from hydrogen production research. Beyond that, I attempt to provide a systematic comparison of different quantum kernel regression approaches and show results from real backend runs. This also incorporates demonstrating effects of proper error mitigation techniques.

[1] M. Schuld and N. Killoran. Phys. Rev. Lett. 122, 040504 (2019)

[2] M. Schuld, arXiv:2101.11020v2 (2021)