Resource-efficient Quantum Neuron for Quantum Neural Networks

Research output: Contribution to book/conference proceedings/anthology/reportConference contributionContributedpeer-review

Abstract

Classical neural networks (CNNs) provide wide applications in communication technology, ranging from enhancement of key performance indicators (KPIs) of communication protocols to utilities in intelligent networks. Quantum neural networks (QNNs) have advantages over CNNs for applications in future communication networks. This article surveys existing methods for implementing QNNs and suggests a novel resource-friendly quantum system with enhanced activation for better performance in QNNs. Our approach is motivated by a principle of neuronal gain control established in neuroscience and was implemented on a quantum computer using IBM's qiskit. Potential implications of such a model for communication technology are discussed.

Details

Original languageEnglish
Title of host publication2023 IEEE Globecom Workshops, GC Wkshps 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1045-1050
Number of pages6
ISBN (electronic)979-8-3503-7021-8
Publication statusPublished - 2023
Peer-reviewedYes

Conference

Title2023 IEEE Globecom Workshops, GC Wkshps 2023
Duration4 - 8 December 2023
CityKuala Lumpur
CountryMalaysia

External IDs

ORCID /0000-0001-8409-5390/work/158767933

Keywords

Keywords

  • CNNs, gain control, non-linear activation functions, QNNs, quantum neuron