TodaysComprehensiveandAuthoritativeGuidetoAugmentedRealityByoverlayingcomputer-generatedinformationontherealworld,augmentedreality(AR)amplifieshumanperceptionandcognitioninremarkableways.Workinginthisfast-growingfieldrequiresknowledgeofmultipledisciplines,includingcomputervision,computergraphics,andhuman-computerinteraction.AugmentedReality:PrinciplesandPracticeintegratesallthisknowledgeintoasingle-sourcereference,presentingtodaysmostsignificantworkwithscrupulousaccuracy.PioneeringresearchersDieterSchmalstiegandTobiasHollerercarefullybalanceprinciplesandpractice,illuminatingARfromtechnical,methodological,anduserperspectives.CoverageincludesDisplays:head-mounted,handheld,projective,auditory,andhapticTracking/sensing,includingphysicalprinciples,sensorfusion,andreal-timecomputervisionCalibration/registration,ensuringrepeatable,accurate,coherentbehaviorSeamlessblendingofrealandvirtualobjectsVisualizationtoenhanceintuitiveunderstandingInteractionfromsituatedbrowsingtofull3DinteractionModelingnewgeometriccontentAuthoringARpresentationsanddatabasesArchitectingARsystemswithreal-time,multimedia,anddistributedelementsThisguideisindispensableforanyoneinterestedinAR,includingdevelopers,engineers,students,instructors,researchers,andserioushobbyists.

Physical-layersecurityanalysisofaquantum-noiserandomizedcipherbasedonthewire-tapchannelmodel

本文主要讨论ApacheSpark的设计与实现，重点关注其设计思想、运行原理、实现架构及性能调优，附带讨论与HadoopMapReduce在设计与实现上的区别。
不喜欢将该文档称之为“源码分析”，因为本文的主要目的不是去解读实现代码，而是尽量有逻辑地，从设计与实现原理的角度，来理解job从产生到执行完成的整个过程，进而去理解整个系统。
讨论系统的设计与实现有很多方法，本文选择问题驱动的方式，一开始引入问题，然后分问题逐步深入。
从一个典型的job例子入手，逐渐讨论job生成及执行过程中所需要的系统功能支持，然后有选择地深入讨论一些功能模块的设计原理与实现方式。
也许这样的方式比一开始就分模块讨论更有主线。
本文档面向的是希望对Spark设计与实现机制，以及大数据分布式处理框架深入了解的Geeks。
因为Spark社区很活跃，更新速度很快，本文档也会尽量保持同步，文档号的命名与Spark版本一致，只是多了一位，最后一位表示文档的版本号。
由于技术水平、实验条件、经验等限制，当前只讨论Sparkcorestandalone版本中的核心功能，而不是全部功能。
诚邀各位小伙伴们加入进来，丰富和完善文档。
好久没有写这么完整的文档了，上次写还是三年前在学Ng的ML课程的时候，当年好有激情啊。
这次的撰写花了20+days，从暑假写到现在，大部分时间花在debug、画图和琢磨怎么写上，希望文档能对大家和自己都有所帮助。
内容本文档首先讨论job如何生成，然后讨论怎么执行，最后讨论系统相关的功能特性。
具体内容如下：Overview总体介绍Joblogicalplan介绍job的逻辑执行图（数据依赖图）Jobphysicalplan介绍job的物理执行图Shuffledetails介绍shuffle过程Architecture介绍系统模块如何协调完成整个job的执行CacheandCheckpoint介绍cache和checkpoint功能Broadcast介绍broadcast功能JobScheduling

Thisbookwaswrittenusingatutorialapproachandisintendedtoappealtoabroadaudience.Thebookcoversthefundamentalsofshort-rangewirelessnetworkingusingZigBee™andIEEE802.15.4™standards.TheZigBeeandIEEE802.15.4standardsarecoveredwiththesamelevelofdetail.Inadditiontotechnicaldetails,thebookcontainshigh-leveloverviewsthatwouldbeinformativeforareaderwhoisonlyinterestedinunderstandingthegeneralconceptofZigBeewirelessnetworking.ThisbookprovidesthebigpictureofZigBeewirelessnetworking,fromtheradiofrequency(RF)andphysicallayerconsiderationsuptotheapplicationlayerdetails.Consideringthemultidisciplinarynatureofthisbook,therequiredbackgroundmaterialsarealsoprovided.

Writtenbypioneersoftheconcept,thisisthefirstcompleteguidetothephysicalandengineeringprinciplesofMassiveMIMO.Assumingonlyabasicbackgroundincommunicationsandstatisticalsignalprocessing,itwillguidereadersthroughkeytopicssuchaspropagationmodels,channelmodeling,andmulti-cellperformanceanalyses.Theauthors’uniquecapacity-boundapproachwillenablereaderstocarryoutmoreeffectivesystemperformanceanalysisanddevelopadvancedMassiveMIMOtechniquesandalgorithms.Numerouscasestudies,aswellasproblemsetsandsolutionsaccompanyingthebookonline,willhelpreadersputknowledgeintopracticeandacquiretheskillsetneededtodesignandanalyzecomplexwirelesscommunicationsystems.Whetheryouareagraduatestudent,researcher,orindustryprofessionalworkinginthefieldofwirelesscommunications,thiswillbeanindispensableguideforyearstocome.

Thisbookcoversalgorithmicandhardwareimplementationtechniquestoenableembeddeddeeplearning.Theauthorsdescribesynergeticdesignapproachesontheapplication-,algorithmic-,computerarchitecture-,andcircuit-levelthatwillhelpinachievingthegoalofreducingthecomputationalcostofdeeplearningalgorithms.Theimpactofthesetechniquesisdisplayedinfoursiliconprototypesforembeddeddeeplearning.Givesawideoverviewofaseriesofeffectivesolutionsforenergy-efficientneuralnetworksonbatteryconstrainedwearabledevices;Discussestheoptimizationofneuralnetworksforembeddeddeploymentonalllevelsofthedesignhierarchy–applications,algorithms,hardwarearchitectures,andcircuits–supportedbyrealsiliconprototypes;ElaboratesonhowtodesignefficientConvolutionalNeuralNetworkprocessors,exploitingparallelismanddata-reuse,sparseoperations,andlow-precisioncomputations;Supportstheintroducedtheoryanddesignconceptsbyfourrealsiliconprototypes.Thephysicalrealization’simplementationandachievedperformancesarediscussedelaboratelytoillustratedandhighlighttheintroducedcross-layerdesignconcepts.

OFDMA系统分析与设计的国外经典教材Chapter1IntroductiontoOFDMandOFDMAChapter2CharacterizationoftheMobileWirelessChannelChapter3FundamentalsofDigitalCommunicationsandNetworkingChapter4FundamentalsofOFDMandOFDMA:TransceiverStructureChapter5PhysicalLayer:TimeandFrequencyChapter6PhysicalLayer:SpatialTechniquesChapter7MediumAccessControl:ArchitectureandDataPlaneChapter8MediumAccessControl:LowerControlPlaneChapter9MediumAccessControl:UpperControlPlaneChapter10QualityofService(QoS)Chapter11SecurityFundamentalsChapter12SecurityFunctionsChapter13RFSystemDesign:CoverageChapter14RFSystemDesign:Capacity

Despitetheincreasinguseofcomputers,thebasicneedformathematicaltablescontinues.Tablesserveavitalroleinpreliminarysurveysofproblemsbeforeprogrammingformachineoperation,andtheyareindispensabletothousandsofengineersandscientistswithoutaccesstomachines.Becauseofautomaticcomputers,however,andbecauseofrecentscientificadvances,agreatervarietyoffunctionsandahigheraccuracyoftabulationthanhavebeenavailableuntilnowarerequired.In1954,aconferenceonmathematicaltables,sponsoredbyM.I.T.andtheNationalScienceFoundation,mettodiscussamodernizationandextensionofJahnkeandEmde'sclassicaltablesoffunctions.Thisvolume,published10yearslaterbytheU.S.DepartmentofCommerce,istheresult.Designedtoincludeamaximumofinformationandtomeettheneedsofscientistsinallfields,itisamonumentalpieceofwork,acomprehensiveandself-containedsummaryofthemathematicalfunctionsthatariseinphysicalandengineeringproblems.Thebookcontains29setsoftables,sometoashighas20places:mathematicalconstants;physicalconstantsandconversionfactors(6tables);exponentialintegralandrelatedfunctions(7);errorfunctionandFresnelintegrals(12);Besselfunctionsofinteger(12)andfractional(13)order;integralsofBesselfunctions(2);Struveandrelatedfunctions(2);confluenthypergeometricfunctions(2);Coulombwavefunctions(2);hypergeometricfunctions;Jacobianellipticandthetafunctions(2);ellipticintegrals{9);Weierstrassellipticandrelatedfunctions;paraboliccylinderfunctions{3);Mathieufunctions(2);spheroidalwavefunctions(5);orthogonalpolynomials(13);combinatorialanalysis(9);numericalinterpolation,differentiationandintegration(11);probabilityfunctions(ll);scalesofnotation(6);miscellaneousfunctions(9);Laplacetransforms(2);andothers.Eachofthesesectionsisprefacedbyalistofrelatedformulasandgraph

Motionplanningisthecentralthemeofthisbook.Itcanbelooselystatedasfollows:Howcanarobotdecidewhatmotionstoperforminordertoachievegoalarrangementsofphysicalobjects?Thiscapabilityiseminentlynecessarysince,bydefinition,arobotaccomplishestasksbymovingintherealworld.

IEEE802.3by™StandardforMediaAccessControlParameters,PhysicalLayersandManagementParametersfor25Gb/sOperation

在日常工作中，钉钉打卡成了我生活中不可或缺的一部分。然而，有时候这个看似简单的任务却给我带来了不少烦恼。 每天早晚，我总是得牢记打开钉钉应用，点击"工作台"，再找到"考勤打卡"进行签到。有时候因为工作忙碌，会忘记打卡，导致考勤异常，影响当月的工作评价。而且，由于我使用的是苹果手机，有时候系统更新后，钉钉的某些功能会出现异常，使得打卡变得更加麻烦。 另外，我的家人使用的是安卓手机，他们也经常抱怨钉钉打卡的繁琐。尤其是对于那些不太熟悉手机操作的长辈来说，每次打卡都是一次挑战。他们总是担心自己会操作失误，导致打卡失败。 为了解决这些烦恼，我开始思考是否可以通过编写一个全自动化脚本来实现钉钉打卡。经过一段时间的摸索和学习，我终于成功编写出了一个适用于苹果和安卓系统的钉钉打卡脚本。