简介：

简介：连接神经原的几何距离的分发是在大脑位于神经网络下面的一个实际因素。它能影响在地面的动态性质铺平的大脑。Karbowski导出被实验还没验证了的幂定律腐烂分布。在这个工作，我们与一个现象学的模型一起用模拟检查它的有效性。基于在里面vitro在由Ito的文化容器的神经网络的二维的开发，我们匹配触处数字浸透时间为开发进程获得合适的参数，然后在如此的条件下面决定在连接神经原之间的距离的分发。我们的模拟获得清楚的指数的分布而不是幂定律，它显示那个Karbowskis结论是无效的，至少为盒子在vitro在二维的文化容器的神经网络开发。

简介：Wepresentthesolidmodeleditdistance(SMED),apowerfulandflexibleparadigmforexploitingshapesimilaritiesamongstCADmodels.ItisdesignedtomeasurethemagnitudeofdistortionsbetweentwoCADmodelsinboundaryrepresentation(B-rep).Wegivetheformaldefinitionbyanalogywithgrapheditdistance,oneofthemostpopulargraphmatchingmethods.Toavoidtheexpensivecomputationalcostpotentiallycausedbyexactcomputation,anapproximateprocedurebasedonthealignmentoflocalstructuresetsisprovidedinaddition.Inordertoverifytheflexibility,wemakeintensiveinvestigationsonthreetypicalapplicationsinmanufacturingindustry,anddescribehowourmethodcanbeadaptedtomeetthevariousrequirements.Furthermore,amultilevelmethodisproposedtomakefurtherimprovementsofthepresentedalgorithmonbotheffectivenessandefficiency,inwhichthemodelsarehierarchicallysegmentedintotheconfigurationsoffeatures.ExperimentresultsshowthatSMEDservesasareasonablemeasurementofshapesimilarityforCADmodels,andtheproposedapproachprovidesremarkableperformanceonareal-worldCADmodeldatabase.

简介：Timesynchronizationtechniques,especiallyonthepulsepersecond(PPS)temporalbasis,haveattractedgrowingresearchinterestsinrecentyears.Inthispaper,wehaveproposedandexperimentallydemonstratedahigh-precisiontwo-waytimetransfer(TWTT)systemtorealizelong-distancedisseminationof1PPSsignalgeneratedbyahydrogenmaser.Adense-wavelength-division-multiplexing(DWDM)systemandbi-directionalerbium-dopedfiberamplifiers(Bi-EDFAs)havealsobeenadoptedtosuppresstheimpactofRayleighbackscatteringandoptimizethesignaltonoiseratio(SNR)aswell.Wehavetheoreticallyanalyzedthesystematicdelayindetail.Theultimaterootmeansquare(RMS)variationoftimesynchronizationaccuracyissub-26psandthetimedeviationcanbereducedtoaslowas1.2psat100sand0.253psat12000s,respectively.