Bonn 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
QI: Fachverband Quanteninformation
QI 2: Quantum Machine Learning I
QI 2.6: Talk
Monday, March 10, 2025, 12:15–12:30, HS VIII
Investigating the Quantum Circuit Born Machine — •Michael Krebsbach1, Florentin Reiter1, Ali Abedi2, Hagen-Henrik Kowalski2, and Thomas Wellens1 — 1Fraunhofer IAF, Tullastraße 72, 79108 Freiburg — 2Bundesdruckerei GmbH, Kommandantenstraße 18, 10969 Berlin
The Quantum Circuit Born Machine (QCBM) is a generative quantum machine learning algorithm that can be used to synthetically extend a dataset that is expensive or otherwise difficult to enlarge. This is achieved by training a parameterized quantum circuit to encode the data distribution p(x) in its output state |ψ⟩ ≈ 1/√N ∑x p(x) |x⟩. Measuring | ψ ⟩ in the computational basis allows to efficiently sample new data points from the distribution.
In this talk, we present our investigation of the trainability and generalization properties of QCBMs. We discuss how the type of data can affect the trainability, and show how it can be improved using several simple techniques. Lastly, we outline how QCBMs could be extended to not only solve generative tasks, but also classification problems.
Keywords: Quantum Ciruit Born Machine; Generative Quantum Machine Learning; Synthetic Data Generation; QCBM; Parameterized Quantum Circuits