人工智能08不确定性.pptx
- 文档编号:15118955
- 上传时间:2023-06-30
- 格式:PPTX
- 页数:54
- 大小:1.21MB
人工智能08不确定性.pptx
《人工智能08不确定性.pptx》由会员分享,可在线阅读,更多相关《人工智能08不确定性.pptx(54页珍藏版)》请在冰点文库上搜索。
手写数字识别,文本分类,图像分割,第八章Uncertainty不确定性,对应教材第13章,本章大纲,Uncertainty不确定性Probability概率SyntaxandSemantics语法与语义Inference推理IndependenceandBayesRule独立性及贝叶斯法则,不确定性,智能体几乎从来无法了解关于其环境的全部事实。
因此其必须在不确定的环境下行动。
概率推理得到了某一证据,那么有多大的几率结论为真?
例如:
我颈部痛;我得脑膜炎的可能有多大?
不确定性,假如有如下规则:
iftoothache(牙疼)then原因是cavity(牙齿有洞)但并不是所有牙疼的病人都是因为牙齿有洞,所以我们可以建立如下规则:
iftoothacheandgum-disease(牙龈疾病)andfilling(补牙)and.thenproblem=cavity以上规则是复杂的;更好的方法:
iftoothachethenproblemiscavitywith0.8probabilityorP(cavity|toothache)=0.8theprobabilityofcavityis0.8giventoothacheisobserved,不确定性,LetactionAt=离起飞时间提前t分钟动身去机场At会使我准时到达机场吗?
Problems:
1.partialobservability/部分可观察性(roadstate,otherdriversplans)2.noisysensors(trafficreports)3.行动结果的不确定性(flattire,etc.)4.immensecomplexityofmodelingandpredictingtraffic因此一个纯粹的逻辑描述方法:
1.risksfalsehood(错误风险):
“A25willgetmethereontime”,or2.leadstoconclusionsthataretooweakfordecisionmaking:
“A25willgetmethereontimeiftheresnoaccidentonthebridgeanditdoesntrainandmytiresremainintactetcetc.”(A1440mightreasonablybesaidtogetmethereontimebutIdhavetostayovernightintheairport),世界与模型中的不确定性,Trueuncertainty:
rulesareprobabilisticinnature掷骰子,抛硬币惰性:
把所有意外的规则都列举出来是很困难的花费太多时间来确定所有的相关因素这些规则过于繁杂而难以使用理论的无知:
某些领域中还没有完整的理论(e.g.,medicaldiagnosis)实践的无知:
掌握了所有规则但是并不是所有的相关信息都能被收集到,处理不确定性的方法,概率理论作为一种正式的方法for:
不确定知识的表示和推理命题中的模型信度(event,conclusion,diagnosis,etc.)给定可获得的证据,A25willgetmethereontimewithprobability0.04概率是不确定性的语言现代AI的中心支柱,Probability概率,概率理论提供了一种方法以概括来自我们的惰性和无知的不确定性。
ProbabilisticassertionssummarizeeffectsofLaziness(惰性):
failuretoenumerateexceptions(例外),qualifications(条件),etc.Ignorance(理论的无知):
lackofrelevantfacts,initialconditions,etc.Subjectiveprobability(主观概率):
Probabilitiesrelatepropositions(命题)toagentsownstateofknowledgee.g.,P(A25|noreportedaccidents)=0.06Thesearenotassertions(断言)abouttheworld命题的概率随着新证据的发现而改变:
e.g.,P(A25|noreportedaccidents,5a.m.)=0.15,不确定条件下的决策,假设下述概率是真的:
P(A25getsmethereontime|)=0.04P(A90getsmethereontime|)=0.70P(A120getsmethereontime|)=0.95P(A1440getsmethereontime|)=0.9999Whichactiontochoose?
Dependsonmypreferences(偏好)formissingflightvs.timespentwaiting,etc.Utilitytheory(效用理论)用来对偏好进行表示和推理Decisiontheory=probabilitytheory+utilitytheory决策理论=概率理论+效用理论,Syntax语法,基本元素:
randomvariable(随机变量)Arandomvariableissomeaspectoftheworldaboutwhichwe(may)haveuncertainty通常大写e.g.,Cavity,Weather,Temperature类似于命题逻辑:
未知世界被随机变量的赋值所定义Booleanrandomvariables(布尔随机变量)e.g.,Cavity(牙洞)(doIhaveacavity?
)Discreterandomvariables(离散随机变量)e.g.,Weatherisoneof定义域mustbeexhaustive(穷尽的)andmutuallyexclusive(互斥的)Continuousrandomvariables(连续随机变量)e.g.,Temp=21.6;alsoallow,e.g.,Temp22.0,Syntax,Elementaryproposition(命题)constructedbyassignmentofavaluetoarandomvariable:
e.g.,Weather=sunny,Cavity=false(简写为cavity)Complexpropositionsformedfromelementarypropositionsandstandardlogicalconnectivese.g.,Weather=sunnyCavity=false,Syntax,Atomicevent:
Acompletespecificationofthestateoftheworldaboutwhichtheagentisuncertain原子事件:
对智能体无法确定的世界状态的一个完整的详细描述。
E.g.,iftheworldconsistsofonlytwoBooleanvariablesCavityandToothache,thenthereare4distinctatomicevents:
Cavity=falseToothache=falseCavity=falseToothache=trueCavity=trueToothache=falseCavity=trueToothache=trueAtomiceventsaremutuallyexclusiveandexhaustive穷尽和互斥,概率公理,对任意命题A,B0P(A)1P(true)=1andP(false)=0P(AB)=P(A)+P(B)-P(AB),Priorprobability(先验概率),Priororunconditionalprobabilities(无条件概率)ofpropositions在没有任何其它信息存在的情况下关于命题的信度e.g.,P(Cavity=true)=0.1andP(Weather=sunny)=0.72correspondtobeliefpriortoarrivalofany(new)evidenceProbabilitydistributiongivesvaluesforallpossibleassignments:
概率分布给出一个随机变量所有可能取值的概率P(Weather)=(normalized(归一化的),i.e.,sumsto1)Jointprobabilitydistributionforasetofrandomvariablesgivestheprobabilityofeveryatomiceventonthoserandomvariables(i.e.,everysamplepoint)联合概率分布给出一个随机变量集的值的全部组合的概率P(Weather,Cavity)=a42matrixofvalues:
Everyquestionaboutadomaincanbeansweredbythejointdistributionbecauseeveryeventisasumofsamplepoints,连续变量的概率,Expressdistributionasaparameterized(参数化的)functionofvalue:
P(X=x)=U18,26(x)=uniform(均匀分布)densitybetween18and26,连续变量的概率,MarginalDistributions(边缘概率分布),Marginaldistributionsaresub-tableswhicheliminatevariablesMarginalization(summingout):
Combinecollapsedrowsbyadding,Conditionalprobability(条件概率),Conditionalorposteriorprobabilities(后验概率)P(a|b)证据累积过程的形式化和发现新证据后的概率更新当一个命题为真的条件下,指定命题的概率e.g.,P(cavity|toothache)=0.8i.e.,鉴于牙疼是已知证据(Notationforconditionaldistributions(条件概率分布):
P(cavity|toothache)=asinglenumberP(Cavity,Toothache)=2x2tablesummingto1P(Cavity|Toothache)=2-elementvectorof2-elementvectorsIfweknowmore,e.g.,cavityisalsogiven,thenwehaveP(cavity|toothache,cavity)=1新证据可能是不相关的,可以简化,e.g.,P(cavity|toothache,sunny)=P(cavity|toothache)=0.8,条件概率,定义条件概率为:
P(a|b)=P(ab)/P(b)ifP(b)0Productrule(乘法规则)givesanalternativeformulation:
P(ab)=P(a|b)P(b)=P(b|a)P(a)Ageneralversionholdsforwholedistributions,e.g.,P(Weather,Cavity)=P(Weather|Cavity)P(Cavity)(Viewasasetof42equations,notmatrixmultiplication)Chainrule(链式法则)isderivedbysuccessiveapplicationofproductrule:
条件概率,条件概率跟标准概率一样,forexample:
0=P(a|e)=1conditionalprobabilitiesarebetween0and1inclusiveP(a1|e)+P(a2|e)+.+P(ak|e)=1conditionalprobabilitiessumto1wherea1,akareallvaluesinthedomainofrandomvariableAP(a|e)=1-P(a|e)negationforconditionalprobabilities,通过枚举的推理,Startwiththejointprobabilitydistribution(全联合概率分布):
Foranyproposition,sumtheatomiceventswhereitistrue:
一个命题的概率等于所有当它为真时的原子事件的概率和,通过枚举的推理,Startwiththejointprobabilitydistribution(全联合概率分布):
Foranyproposition,sumtheatomiceventswhereitistrue:
一个命题的概率等于所有当它为真时的原子事件的概率和,通过枚举的推理,Startwiththejointprobabilitydistribution(全联合概率分布):
Foranyproposition,sumtheatomiceventswhereitistrue:
一个命题的概率等于所有当它为真时的原子事件的概率和,通过枚举的推理,Startwiththejointprobabilitydistribution(全联合概率分布):
Normalization(归一化),Denominator(分母)canbeviewedasanormalizationconstantP(Cavity|toothache)=P(Cavity,toothache)=P(Cavity,toothache,catch)+P(Cavity,toothache,catch)=+=Generalidea:
computedistributiononqueryvariablebyfixingevidencevariables(证据变量)andsummingoverhiddenvariables(未观测变量),通过枚举的推理,Typically,weareinterestedintheposteriorjointdistributionofthequeryvariables(查询变量)Ygivenspecificvaluesefortheevidencevariables(证据变量)ELetthehiddenvariables(未观测变量)beH=X-YEThentherequiredsummationofjointentriesisdonebysummingoutthehiddenvariables:
P(Y|E=e)=P(Y,E=e)=hP(Y,E=e,H=h)ThetermsinthesummationarejointentriesbecauseY,EandHtogetherexhaustthesetofrandomvariables(Y,E,H构成了域中所有变量的完整集合)Obviousproblems:
1.Worst-casetimecomplexityO(dn)wheredisthelargestarity2.SpacecomplexityO(dn)tostorethejointdistribution3.HowtofindthenumbersforO(dn)entries?
Independence(独立性),AandBareindependentiffP(A|B)=P(A)orP(B|A)=P(B)orP(A,B)=P(A)P(B)E.g:
rollof2die:
P(1,3)=1/6*1/6=1/36,P(Toothache,Catch,Cavity,Weather)=P(Toothache,Catch,Cavity)P(Weather)32entriesreducedto12;fornindependentbiasedcoins,O(2n)O(n)Absoluteindependencepowerfulbutrare绝对独立强大但罕见Dentistry(牙科领域)isalargefieldwithhundredsofvariables,noneofwhichareindependent.Whattodo?
独立的滥用,天真的数学笑话:
一个著名统计学家永远不会坐飞机旅行,因为他研究了航空旅行和估计,任何给定的航班上有炸弹的可能性是一百万分之一,他不准备接受这些可能性。
有一天,一位同时在远离家乡的会议上遇到他。
“你怎么到这里的?
坐火车吗?
”“不,我飞过来的”“Whataboutthepossibilityofabomb?
”“Well,Ibeganthinkingthatiftheoddsofonebombare1:
million,thentheoddsoftwobombsare(1/1,000,000)x(1/1,000,000).Thisisavery,verysmallprobability,whichIcanaccept.SonowIbringmyownbombalong!
”,Conditionalindependence条件独立性,Randomvariablescanbedependent,butconditionallyindependentExample:
YourhousehasanalarmNeighborJohnwillcallwhenhehearsthealarmNeighborMarywillcallwhenshehearsthealarmAssumeJohnandMarydonttalktoeachotherIsJohnCallindependentofMaryCall?
NoIfJohncalled,itislikelythealarmwentoff,whichincreasestheprobabilityofMarycallingP(MaryCall|JohnCall)P(MaryCall),条件独立性,But,ifweknowthestatusofthealarm,JohnCallwillnotaffectwhetherornotMarycallsP(MaryCall|Alarm,JohnCall)=P(MaryCall|Alarm)WesayJohnCallandMaryCallareconditionallyindependentgivenAlarmIngeneral,“AandBareconditionallyindependentgivenC”means:
P(A|B,C)=P(A|C)P(B|A,C)=P(B|C)P(A,B|C)=P(A|C)P(B|C),条件独立性,P(Toothache,Cavity,Catch)has23-1=7independententries专业领域知识:
Cavitydirectlycausestoothacheandprobe-catches.IfIhaveacavity,theprobabilitythattheprobecatchesinitdoesntdependonwhetherIhaveatoothache:
(1)P(catch|toothache,cavity)=P(catch|cavity)ThesameindependenceholdsifIhaventgotacavity:
(2)P(catch|toothache,cavity)=P(catch|cavity)CatchisconditionallyindependentofToothachegivenCavity:
P(Catch|Toothache,Cavity)=P(Catch|Cavity)Equivalentstatements:
P(Toothache|Catch,Cavity)=P(Toothache|Cavity)P(Toothache,Catch|Cavity)=P(Toothache|Cavity)P(Catch|Cavity),条件独立性,Writeoutfulljointdistributionusingchainrule:
P(Toothache,Catch,Cavity)=P(Toothache|Catch,Cavity)P(Catch,Cavity)=P(Toothache|Catch,Cavity)P(Catch|Cavity)P(Cavity)=P(Toothache|Cavity)P(Catch|Cavity)P(Cavity)I.e.,2+2+1=5independentnumbersInmostcases,theuseofconditionalindependencereducesthesizeoftherepresentationofthejointdistributionfromexponentialinntolinearinn.在大多数情况下,使用条件独立性能将全联合概率的表示由n的指数关系减为n的线性关系。
Conditionalindependenceisourmostbasicandrobustformofknowledgeaboutuncertainenvironments.,BayesRule(贝叶斯法则),BayesRule(贝叶斯法则),乘法原则Bayesrule:
orindistributionform为什么该法则非常有用?
将条件倒转通常一个条件是复杂的,一个是简单的许多系统的基础(e.g.语音识别)现代AI基础!
BayesRule(贝叶斯法则),Usefulforassessingdiagnosticprobability(诊断概率)fromcausalprobability(因果概率):
E.g.,letMbemeningitis(脑膜炎),Sbestiffneck(脖子僵硬):
Note:
脑膜炎的后验概率依然非常小!
Note:
依然要先
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- 人工智能 08 不确定性