2024 全球6G技术大会 -10.0E Edge-Native Intelligence White Paper.docx

1/123

Abstract

Edgecomputing,asakeytechnologyofthenextgenerationofradioaccess

networks(RAN),hasdriventhedecentralizationofnetworksandcomputingfacilities.Edgeserversclosertouserterminalscansignificantlyreduceservicelatencyandcope

withemergingnewscenarios.Simultaneously,therapiddevelopmentofartificial

intelligence(AI)playsasignificantroleinenhancingtheperformanceofedge

computing,aidingedgedevicesincopingwiththerapidlyincreasingdataontheedge.

Therefore,combiningthelocalcomputingcapabilityofedgedatawiththestrong

computingcapabilitiesofAI,knownasedgeintelligence,canenhancethedata

processingcapabilitiesontheedge,improvetheoverallperformanceofwireless

communicationsystems,andenhanceuserserviceexperiences.Edgeintelligenceisahotandrapidlydevelopingfieldinrecentyears,andthiswhitepaperaimstoanalyze

thecurrentresearchprogressinedgeintelligence.Itmainlyincludes:

(1)6GEdgeIntelligenceNetworksandInfrastructure:Firstly,theedge-native

intelligentarchitecturefor6Gnetworksisanalyzed.Then,theedgeintelligence

computinginfrastructureisintroduced,includingedgeintelligenthardwareandcloudplatforms.Finally,theedgeintelligencenetworkinfrastructureisdescribed,including

theedgeintelligenceaccessnetworkandcorenetwork.

(2)KeyTechnologiesofEdge-NativeIntelligence:Itisintroducedfromthe

aspectsofmodellightweight,edge-cloudcollaborativeintelligence,edgeintelligentdeployment,anddeepedgenodes.Edgeintelligenceinwirelessfederatedlearningis

alsoexplainedindetail,includingmodelsparsificationandmodelquantizationin

federatedlearning.

(3)ApplicationsofEdge-NativeIntelligence:Typicalapplicationsofedge-native

intelligenceareanalyzed,suchassmarttransportation,smartmanufacturing,and

smartenergysaving.

2/123

Contents

1.Introduction 3

1.1Background 3

1.2Overview