Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz
AKPIK 4: Postersession AKPIK
AKPIK 4.1: Poster
Mittwoch, 27. März 2019, 18:00–18:30, C.A.R.L. Foyer EG
Towards Realizing Machine Learning Using Spin Qubits based on GaAs — •Zheng Zeng1, Yulin Hu2, Beata Kardynał1, and Anke Schmeink2 — 1Peter Grünberg Institute (PGI-9), Forschungszentrum Jülich, D-52425 Jülich, Germany — 2Theoretische Informationstechnik, RWTH Aachen University, D-52056 Aachen, Germany
Machine learning (ML) has been a powerful tool for executing advanced inference tasks. Recently, implementing ML on a powerful quantum computer becomes an attractive research filed as it has been shown that quantum computers exhibit square-root and even exponential speedups over classical computers in some machine learning methods based on quantum algorithms. Different from classical computers, quantum computers are able to process information using effects like quantum coherence and entanglement. Therefore, as a basic unit of quantum computation, a qubit has a high information capacity, i.e., it could carry much more information than a classical bit. In this work, we are motivated to investigate a possibility to improve the performance of ML programs by utilizing the high information capacity characteristic of qubits. In particular, we provide a case study to realize a neural network via a set of a few spin qubits. We propose a protocol to encode the training set of the neural network by manipulating the spin qubits on the Bloch sphere using radio frequency (RF) pulses, and finally evaluate the performance of the neural network.