麦肯锡《信息技术》15 期Word格式文档下载.docx

麦肯锡《信息技术》15 期Word格式文档下载.docx

《麦肯锡《信息技术》15 期Word格式文档下载.docx》由会员分享，可在线阅读，更多相关《麦肯锡《信息技术》15 期Word格式文档下载.docx（70页珍藏版）》请在冰豆网上搜索。

New-Tech（科技前沿）3

BigDataAnalysisIsChangingtheNatureofSportsScience3

大数据分析正在改变体育科学本质5

2016isturningouttobeanamazingyearforaugmentedreality7

2016年增强现实技术蓬发展8

AMDdoublesdownonVRwithSulonQ,LiquidVR10

超微半导体公司携SulonQ,LiquidVR双重来袭12

Essay:

WilltheFirstAmendmentsurvivetheinformationage?

13

《第一修正案》在信息时代能存活下去吗？

15

Unlockingthesecretsofthebrain'

sintelligencetodevelopsmartertechnologies17

解锁大脑的智力秘密以开发更加智能的技术18

NotesfromSXSWInteractive:

Moore’sLaw,software,anddisruption20

“南偏西南”互动媒体大会关键词：

摩尔定律、软件和介入22

ComputerSoftware&

Application（计算机应用与软件）23

LawrenceLivermoreandIBMcollaboratetobuildnewbrain-inspiredsupercomputer23

劳伦斯•利弗莫尔实验室和IBM合作创建新型脑启发超级计算机24

Toolchainforreal-timeprogramming25

实时编程工具链26

SeagateunveilsPCIex16SSDwith10GB/sbandwidthatOpenComputeSummit26

希捷科技于开源计算峰会首次发布秒速10GB的PCIex16固态驱动器27

5bestadd-onsfortheRaspberryPi328

树莓派迷你小电脑的5个顶级附加组件29

MicrosoftSaysMaverickChatbotTayForeshadowstheFutureofComputing30

微软表示特立独行的聊天机器人Tay预示着计算机技术的发展前景30

New'

machineunlearning'

techniquewipesoutunwanteddataquicklyandcompletely31

新型“机器免学”技术可快速彻底清除无用数据33

Instrumenterror:

AMD,FCAT,andAshesoftheSingularitybenchmarks35

仪器误差：

AND、FCAT以及奇点灰烬的基准39

Communication（通信）42

Proposedbillwouldblockanonymoussaleof‘burner’phonesinUS42

美国提案将阻止匿名销售一次性手机43

It’stimetoholdmanufacturersresponsibleforAndroidvulnerabilities43

是时候让制造商为Android漏洞负责了44

Assmartphonesstall,PCandmobilemanufacturersslugitoutover2-in-1PCs45

随着智能手机暂停，PC和手机厂商在2合1电脑方面决一雌雄46

AppleGoesAllOutOnRivalMicrosoft46

苹果全线紧逼微软47

Intenet（网络）48

FacedwithFCCregulationsonroutercapabilities,TP-Linkblocksopen-sourceupdates48

在联邦通信委员会的严令限制下，TP-Link屏蔽了路由器的开源升级功能50

Researcherssaynewgenerationofransomwareemerging51

研究人员宣告新一代勒索软件出现52

ComcastrollsoutgigabitinAtlanta,usesdatacapstoforcecustomersintocontracts52

康卡斯特公司在亚特兰大推出千兆速，并利用数据限额强迫顾客接受合同53

Thecybersecuritythreat–areweprotectedyet?

54

网络安全的威胁—我们还在被保护着吗？

55

Turkeypasseslong-awaiteddataprotectionlaw56

土耳其通过了期待已久的数据保护法57

NewregulationscouldfurthercloseChina'

sInternet57

新法规的颁布使中国网络更为封闭58

New-Tech（科技前沿）

BigDataAnalysisIsChangingtheNatureofSportsScience

Thebest-sellingbookMoneyballbyMichaelLewischangedthewaypeoplethoughtaboutsport,particularlyforthoseowners,managers,andplayerswiththebiggestvestedinterests.Lewis’sbookhelpedbringaboutarevolutioninwhichplayerperformancewasmeasuredandassessedusinganevidence-basedapproachratherthanatraditiondominatedbyanecdoteandintuition.

Sincethen,sportsscientistshaveattemptedtoreplicatethesuccessofthisapproachinsportssuchasbasketball,soccer,Americanfootball,andsoon.Thisscienceisdrivenbytherelativelynewabilitytogathervastamountsofdataabouttheplayersandtheplaywhilethegameisinprogress.

However,inmanyofthesesports,thecapacitytogatherdatahasnotbeenmatchedbyanabilitytoprocessitinmeaningfulways.Soaninterestingquestioniswhatchallengessportssciencesfaceincrunchingthisdataeffectively.Whataretheopenquestionsinthisrapidlyevolvingfield?

TodaywegetananswerthankstotheworkofJoachimGudmundssonandMichaelHortonattheUniversityofSydneyinAustralia,whohavereviewedthisfieldandlistedtheoutstandingchallengesthatresearchersfaceinmakinganalyticsmeaningful.

Thesportstheseguysconsideraretogetherknownasinvasiongames.Theyallconsistoftwoteamsthatcompeteforpossessionofaballinaconstrainedplayingarea.Eachteamhasthesimultaneousobjectiveofscoringbyputtingtheballintotheopposition’sgoalandalsoofdefendingitsowngoal.Theteamthatscoresthegreatestnumberofgoalsbytheendofthegameisthewinner.

Invasionsportsthatsharethisstructureincludesoccer,basketball,icehockey,fieldhockey,rugby,Australianrulesfootball,Americanfootball,lacrosseandsoon.However,mostofthedatacomesfromgamessuchasprofessionalsoccerandbasketball,whichhavetheresourcestogatherit.

Thisdatagenerallyconsistsofplayerandballtrajectoriesthroughoutthegame,andeventlogsthatdescribeeventssuchaspasses,shots,tackles,andsoonatspecifictimes.“Stateoftheartobjecttrackingsystemsnowproducespatio-temporaltracesofplayertrajectorieswithhighdefinitionandhighfrequency,andthis,inturn,hasfacilitatedavarietyofresearchefforts,acrossmanydisciplines,toextractinsightfromthetrajectories,”sayGudmundssonandHorton.

Thebigchallengeinsportsscienceistousethisdatatogainacompetitiveadvantage,whetherinrealtimeduringthegameortohelpintraining,preparation,orrecruitment.Butwhileresearchershavemadesignificantprogress,therearealsoimportanthurdlesbarringtheway.

Oneofthemostsignificantinvolvesunderstandinghowplayerscandominatepartsofthepitchnearthem.Insportsscience,aplayer’sdominantregionistheregionheorshecanreachbeforeanyotherplayer.AsimplewaytocalculatethisistodrawaVoronoidiagram,whichdividesthepitchintotheregionsclosesttoeachplayer（seediagram）.

Suchadiagramcanbemodifiedwiththehelpofotherinformation,suchastheobservationthatdominantregionstendtobelargerfortheattackingteamthanthedefendingteam.

However,calculatingtheVoronoidiagramforeachplayeronthepitchiscomputationallyexpensive.Nobodyhassuccessfullydoneitinrealtime,evenforRoboCupfootball.

Instead,researcherscalculateadifferentproperty—theregioneachplayercanreachinagiventime—andthenlookforoverlaps,whicharethenresolved.Thisincreasesspeedbyafactorof1,000atacostofa10percentlossinaccuracy.

Buteventhen,thisapproachignoresanumberofcrucialfactors.Perhapsthemostsignificantisthatittakesnoaccountoftheplayers’momentum.Clearly,aplayerinmotioncandominateagreaterregionaheadthanastationaryplayer.

Thiscanleadtocomplexsubdivisionsofthepitch.WhenplayerArunsatanopposingplayerBwhoisstationary,eachmayhavemorethanonedominantregion,andthesemaynotbeconnectedtoeachother.Forexample,playerA’smomentumgivesbetteraccesstosome,butnotall,oftheregionbehindB.

Soanimportantopenprobleminsportsscienceishowtocalculaterealisticdominantregionsinrealtime.

Anotherrelatedchallengeistoworkoutwhetheraplayerisopentoreceiveapass.Thatmeansdeterminingifthereisacertainspeedanddirectionthattheballcanbepassedsothatagivenplayercaninterceptitbeforeanyother.

Thisisobviouslylinkedtotheplayer’sdominantregion.Givenanaccurateideaofwhatthatregionis,it’sstraightforwardtoworkoutastraight-linepassthatfallswithinit.Indeed,that’showthecurrenttoolsthatdothiswork.

Theproblemisthatonlycertaintrajectoriesmeetthecriterionofbeingstraight-linepasses.Anaerialtrajectory,forexample.isnotastraight-linepass.Notoolyetexiststhatcanhandlethese（ormorecomplexmotionsinvolvingthespinoftheball）,andthisisanotheropenprobleminsportsscience

Thenthereisthewaythatoneplayercanputpressureonotherplayersbyclosingdownthespacearoundthem.Howcanthisbemeasuredandincorporatedintomodels?

Anincreasinglyimportantareaofsportinganalysisinvolvesnetworkscience.Thistreatseachplayerasanodeanddrawsalinebetweenthemwhentheballtravelsfromonetotheother.Thishasbeenafruitfulareaofresearchbecauseawiderangeofmathematicaltoolshavealreadybeendevelopedforanalyzingnetworks.

Forexample,itisstraightforwardtoworkoutthemostimportantnodesinthenetworkusingameasureknownascentrality.Insoccer,goalkeepersandforwardshavethelowestcentrality,whiledefendersandmidfieldershavethehighest.

Thesamekindofsciencealsoallowsthenetworktobedividedintoclusters.Sosometeammembersmightonlypasstoeachotherorworkmoreeffectivelytogether.

However,theproblemwithnetworkscienceisthattherearenumerousdifferentwaysofmeasuringcentralityanddeterminingclusters,anditisnotalwaysclearwhyonemethodshouldbepreferredoveranother.Soanotheropenproblemistosystematicallyevaluateandcomparethesedifferentmethodstodeterminetheirutilityandvalue.

Anotherclassofproblemscomefromanalyzinggame-playdata.Forexample,giventhelistofplayertrajectoriesandeventlogsforaperiodduringthegame,isitpossibletodeterminetheteamformation–forexample,4-4-2insoccer–orthetypeofmarkingusedbythedefensiveteam,suchasafull-courtpressorazonedefenseinbasketball?

Thereissomeevidencethatthiscanbedonesomeofthetimeincertainsports.Butmatchingorbeatinghumanperformanceinthisisstillthegoal.

GudmundssonandHortondescribevariousotheropenproblemsandhowideasdevelopedinsportssuchasfootballandbasketballcouldusefullybeappliedinotherinvasionsports,suchashockeyandhandball.

Butperfectingalgorithmsthatcansolvetheseproblemsisonlyhalfthebattle.Thenextstagewillbetoaskhowthesetoolscanhelpimproveperformancebothonandoffthefield.Cantheybeusedasametricofplayerperformanceandvalue?

Cantheydeterminewhetheraplayerwhoissuccessfulononeteamwillbealsobesuccessfulonanother?

Andcantheyworkinrealtimeduringagametohelpcoachesandfansalike?

Therearelikelytobesignificantdevelopmentsinthecomingyears.Clea