国际会议演讲稿

1、国际会议演讲稿自我介绍Thankyou,Mr./Ms.Chair./professorMynameissangqian.Iamveryhonoredtobeheretodooralpresentation.IamaMasterstudentfromHohaiUniversityandIamcurrentlydoingsomeresearchonphysicallayersecurity.Today,Iwouldliketosharewithyousomeofmyresearchonrelayselectionincooperativecommunication.(external/ekˈstərnəl;ɪkˈstərnəl/)内容安排：Mypresentationincludesthesefiveparts.First,somebackgroundinformationaboutthisresearch;Second,systemmodelwehavedone;Third,NN-basedrelayselectionschemewehaveproposedForth,Simulati。

2、onandresultsanalysisAndlast,someconclusionswehavegotP4Partone,introductionFirstly,Iwouldliketogiveyouabitofbackground.Differingfromthetraditionalcryptographictechniquesbasedonsecretkeys,wecanmakeuseofwirelesschannelcharacteristicstoenhancephysicallayersecurity.Cooperativecommunicationhasbeenwidelyrecognizedasaneffectivewaytocombatwirelessfadingandprovidediversitygainwhichisoneoftheresearchhotspots.Machinelearningasanemergingtechnologyhasbeenwidelyappliedinimageprocessing,cancerprediction,stockan。

3、alysisandotherfields.Sowhynottryitinwirelesscommunication?P5:Next,IwanttotalkalittlebitaboutpresentstudyRecentstudiesondeeplearningforwirelesscommunicationsystemshaveproposedalternativeapproachestoenhancecertainpartsoftheconventionalcommunicationsystemsuchasmodulationrecognition、channelencodinganddecoding、channelestimationanddetectionandanautoencoderwhichcanreplacethetotalsystemwithanovelarchitecture【modulationrecognition:AnNNarchitectureformodulationrecognitionthatconsistsofa4-layerNNandtwotwo-。

4、layerNNs。channelencodinganddecoding：AplainDNNarchitectureforchanneldecodingtodecodekbitsmessagesfromNbitsnoisycodewords。channelestimationanddetection：Adense-Netforsymbol-to-symboldetectioncanadoptlongshort-termmemory(LSTM)todetectanestimatedsymbol.Autoencoder：theautoencodercanrepresenttheentirecommunicationsystemandjointlyoptimizethetransmitterandreceiveroveranAWGNchannel.】P6Sowhydidweconductthisresearch?Well,wewanttoexploitthepotentialbenefitsofdeeplearninginenhancingphysicallayersecurityincoop。

5、erative(/kəʊ'ɒpərətɪv/wirelesscommunicationandreducethefeedbackoverheadinlimitedspectrumresoucebyourourproposedscheme.SDE1R2RNR...,1srh,2srh,srNh,1rdh,2rdh,rdNh,1reg,2reg,reNg1ry1rx2ry2rxrNyrNxP8Nowletmemoveontoparttwo-systemmodelHere,youcanseeafigurewhichisasystemmodel.Thisisthesource;thesearetherelaynodesandthisisthedestination,thisistheeavesdropperThewholeprocessofcooperativewirelesscommunicationcanbedividedintotwophases.Inthefirstphase,thesourcebroadcaststhesignaltotheoptimalrelaywhichguaran。

6、teesperfectsecurity.AsshowninFig1,,srihrepresentsafadingcoefficientofthechannelfromthesourcetotherelaynode(iR.)Inthesecondphase,theoptimalrelayforwardsascaledversionofitsreceivedsignaltothedestinationinthepresenceoftheeavesdropper,wheretheoptimalrelayisconsideredtoadoptamplify-and-forward(AF)relayscheme.Inthisfigure,,rdihrepresentsafadingcoefficientofthechannelfromtherelayiRtothedestination,reigrepresentsafadingcoefficientofthechannelfromtherelayiRtotheeavesdropper.P9:Hereyoucanseesomefollowinge。

7、xpressions.Iamnotgoingtowasteourprecioustimeonthelengthyderivation.Iwouldliketoinviteyoutodirectlytakealookattheequationinitsfinalform.Thisistheoptimalindexoftheselectedrelaywiththeconventionalrelayselectionscheme.Amaongthisexpression,,,=max,0sidieiCCCrepresentstheachievablesecrecyrateofsystemmodelwhenthe-thirelayisselected.P11Nowletmemovetopartthree-----NN-basedRelaySelectionHereyoucanseeafigurewhichshowsconventional3-layerneuralnetwork.Itconsistsofinputlayer,hiddenlayer1,hidden(/'hɪdn/)laye。

8、r2andoutputlayer.Neuralnetworkcanlearnfeaturesfromrawdataautomaticallyandadjustparameters（/pəˈræmɪtə(r)z/）flexibly(/'fleksəbli/)suchasweightsandbiases.Incomplex(/'kɒmpleks/)conditions(scenarios(/sɪ'nɑːrɪəʊ/),)Neuralnetworkhaspromisingapplicationsinrelayselectionforseveralreasons.First,thedeepnetworkhassuperior(/suːˈpɪərɪə/)learningabilitydespite(/dɪ'spaɪt/)thecomplexchannelconditions.Second,Neuralnetworkcanhandlelargedatasetsbecauseofdistributed(/dɪ'strɪbjʊtɪd/)andparallel(/'pærəlel/)computing(/。

9、kəm'pjuːtɪŋ/s,whichensurecomputation(/kɒmpjʊ'teɪʃ(ə)n/)speedandprocessingcapacity(/kə'pæsɪtɪ/).Third,variouslibrariesorframeworks,suchasTensorFlow,Theano,andCaffegiveitwideapplicationsInthispaper,theproblemoftherelayselectionismodeledasamulti(/'mʌltɪ/,ao)-classificationproblem.Weadoptsimpleneuralnetwork(NN)toselecttheoptimalrelaytoguaranteesperfectsecrecyperformanceofrelaycooperativecommunicationsystem.(enhancephysicallayersecurity)P12Beforetrainingtheclassificationmodel,weneedtomakesomepreparat。

10、ionfordeeplearningtoacquireatrainingsetandatestingset.First,weneedtoproducerealfeaturevectorforeachexampleaccordingtochannelstateinformation;becausethechannelstateinformationmatricesiscomposedofcomplexnumbersbutfeaturevectorsaregenerallycomposedofrealnumbers.Soweneedtochangecomplexnumbersintorealnumberswithabsolute(/'æbsəluːt/)valueoperation.Moreover,inordertoimprovetheclassificationperformance(precision),itisnecessarytonormalizethefeaturevectors.Second,weneedtodesignkeyperformanceindicator(KPI).。