Artificial Intelligence: A Modern Approach

1127 Pages·2019·21.14 MB·English

by  Stuart Russell, Peter Norvig

ForLoy, Gordon,Lucy, George, andIsaac—S.J.R.
ForKris,Isabella,and Juliet— P.N.
Preface
Artificial Intelligence (AI) is a big field, and this is a big book. We have tried to explore
thefullbreadthofthefield,whichencompasseslogic,probability, andcontinuousmathemat-
ics; perception, reasoning, learning, and action; fairness, trust, social good, and safety; and
applications that range from microelectronic devices torobotic planetary explorers to online
services withbillionsofusers.
The subtitle of this book is “A Modern Approach.” That means we have chosen to tell
the story from a current perspective. We synthesize what is now known into a common
framework, recasting early work using the ideas and terminology that are prevalent today.
Weapologize tothosewhosesubfieldsare,asaresult,lessrecognizable.
New to this edition
Thisedition reflectsthechangesinAIsincethelasteditionin2010:
• We focus more on machine learning rather than hand-crafted knowledge engineering,
duetotheincreased availability ofdata,computing resources, andnewalgorithms.
• Deep learning, probabilistic programming, and multiagent systems receive expanded
coverage,eachwiththeirownchapter.
• The coverage of natural language understanding, robotics, and computer vision has
beenrevisedtoreflecttheimpactofdeeplearning.
• Theroboticschapternowincludes robotsthatinteractwithhumansandtheapplication
ofreinforcement learningtorobotics.
• PreviouslywedefinedthegoalofAIascreating systemsthattrytomaximizeexpected
utility,wherethespecificutilityinformation—the objective—is suppliedbythehuman
designersofthesystem. Nowwenolongerassumethattheobjectiveisfixedandknown
bytheAIsystem;instead, thesystem maybeuncertain aboutthetrueobjectives ofthe
humansonwhosebehalfitoperates. Itmustlearnwhattomaximizeandmustfunction
appropriately evenwhileuncertain abouttheobjective.
• WeincreasecoverageoftheimpactofAIonsociety,includingthevitalissuesofethics,
fairness, trust,andsafety.
• We have moved the exercises from the end of each chapter to an online site. This
allows usto continuously add to, update, and improve the exercises, tomeet the needs
ofinstructors andtoreflectadvances inthefieldandinAI-relatedsoftwaretools.
• Overall, about 25% of the material in the book is brand new. The remaining 75% has
beenlargelyrewrittentopresentamoreunifiedpictureofthefield. 22%ofthecitations
inthiseditionaretoworkspublished after2010.
Overview of the book
The main unifying theme is the idea of an intelligent agent. We define AI as the study of
agents that receive percepts from the environment and perform actions. Each such agent
implements a function that maps percept sequences to actions, and we cover different ways
to represent these functions, such as reactive agents, real-time planners, decision-theoretic
vii
viii Preface
systems, and deep learning systems. We emphasize learning both as a construction method
for competent systems and as a way of extending the reach of the designer into unknown
environments. We treat robotics and vision not as independently defined problems, but as
occurringintheserviceofachievinggoals. Westresstheimportanceofthetaskenvironment
indetermining theappropriate agentdesign.
Our primary aim is to convey the ideas that have emerged over the past seventy years
of AI research and the past two millennia of related work. We have tried to avoid exces-
siveformalityinthepresentation oftheseideas,whileretaining precision. Wehaveincluded
mathematical formulas and pseudocode algorithms to make the key ideas concrete; mathe-
maticalconcepts andnotation aredescribed inAppendix Aandourpseudocode isdescribed
inAppendixB.
This book is primarily intended for use in an undergraduate course or course sequence.
The book has 28 chapters, each requiring about a week’s worth of lectures, so working
through the whole book requires a two-semester sequence. A one-semester course can use
selected chapters to suit the interests of the instructor and students. The book can also be
used in a graduate-level course (perhaps with the addition of some of the primary sources
◮ suggested inthebibliographical notes),orforself-study orasareference.
Throughout the book, important points are marked with a triangle icon in the margin.
Term Wherever a new term is defined, it is also noted in the margin. Subsequent significant uses
oftheterm areinbold, butnotinthemargin. Wehaveincluded acomprehensive index and
anextensivebibliography.
Theonlyprerequisite isfamiliaritywithbasicconceptsofcomputerscience(algorithms,
data structures, complexity) at a sophomore level. Freshman calculus and linear algebra are
usefulforsomeofthetopics.
Online resources
Online resources are available through pearsonhighered.com/cs-resources or at the
book’sWebsite,aima.cs.berkeley.edu.Thereyouwillfind:
• Exercises,programming projects, andresearch projects. Thesearenolongerattheend
of each chapter; they are online only. Within the book, we refer to an online exercise
with a name like “Exercise 6.NARY.” Instructions on the Web site allow you to find
exercisesbynameorbytopic.
• Implementations ofthealgorithmsinthebookinPython,Java,andotherprogramming
languages (currently hostedatgithub.com/aimacode).
• A list of over 1400 schools that have used the book, many with links to online course
materialsandsyllabi.
• Supplementary materialandlinksforstudentsandinstructors.
• Instructions onhowtoreporterrorsinthebook,inthelikelyeventthatsomeexist.
Book cover
The cover depicts the final position from the decisive game 6 of the 1997 chess match in
whichtheprogramDeepBluedefeatedGarryKasparov(playingBlack),makingthisthefirst
time a computer had beaten a world champion in a chess match. Kasparov is shown at the
Preface ix
top. To his right is a pivotal position from the second game of the historic Go match be-
tweenformerworldchampion LeeSedolandDeepMind’s ALPHAGO program. Move37by
ALPHAGO violated centuries ofGoorthodoxy and wasimmediately seen by human experts
as an embarrassing mistake, but it turned out to be a winning move. At top left is an Atlas
humanoid robot built by Boston Dynamics. A depiction of a self-driving car sensing its en-
vironment appears between AdaLovelace, theworld’s firstcomputer programmer, andAlan
Turing, whose fundamental work defined artificial intelligence. At the bottom of the chess
board areaMarsExploration Roverrobot andastatue ofAristotle, whopioneered thestudy
oflogic;hisplanningalgorithmfromDeMotuAnimaliumappearsbehindtheauthors’names.
BehindthechessboardisaprobabilisticprogrammingmodelusedbytheUNComprehensive
Nuclear-Test-BanTreatyOrganizationfordetecting nuclearexplosionsfromseismicsignals.
Acknowledgments
It takes a global village to make a book. Over 600 people read parts of the book and made
suggestions for improvement. Thecomplete list is at aima.cs.berkeley.edu/ack.html;
wearegratefultoallofthem. Wehavespaceheretomentiononlyafewespeciallyimportant
contributors. Firstthecontributing writers:
• JudeaPearl(Section13.5,CausalNetworks);
• VikashMansinghka (Section15.3,ProgramsasProbability Models);
• MichaelWooldridge (Chapter18,Multiagent DecisionMaking);
• IanGoodfellow (Chapter21,DeepLearning);
• JacobDevlinandMei-WingChang(Chapter24,DeepLearningforNaturalLanguage);
• JitendraMalikandDavidForsyth(Chapter25,ComputerVision);
• AncaDragan(Chapter26,Robotics).
Thensomekeyroles:
• CynthiaYeungandMalikaCantor(projectmanagement);
• JulieSussmanandTomGalloway(copyediting andwritingsuggestions);
• OmariStephens(illustrations);
• TracyJohnson(editor);
• ErinAultandRoseKernan(coverandcolorconversion);
• Nalin Chhibber, Sam Goto, Raymond de Lacaze, Ravi Mohan, Ciaran O’Reilly, Amit
Patel,DragomirRadiv,andSamagraSharma(onlinecodedevelopmentandmentoring);
• GoogleSummerofCodestudents (onlinecodedevelopment).
Stuartwouldliketothankhiswife,LoySheflott,forherendlesspatienceandboundless
wisdom. He hopes that Gordon, Lucy, George, and Isaac will soon be reading this book
after they have forgiven him for working so long on it. RUGS (Russell’s Unusual Group of
Students)havebeenunusually helpful, asalways.
Peter would like to thank his parents (Torsten and Gerda) for getting him started, and
hiswife(Kris),children(BellaandJuliet), colleagues, boss, andfriendsforencouraging and
tolerating himthrough thelonghoursofwritingandrewriting.
About the Authors
Stuart Russell was born in 1962 in Portsmouth, England. He received his B.A. with first-
class honours inphysics fromOxfordUniversity in1982, andhisPh.D.incomputer science
from Stanford in 1986. He then joined the faculty of the University of California at Berke-
ley, where he is aprofessor and former chair of computer science, director of the Center for
Human-Compatible AI, and holder of the Smith–Zadeh Chair in Engineering. In 1990, he
received thePresidential YoungInvestigator AwardoftheNational Science Foundation, and
in1995hewascowinneroftheComputersandThoughtAward. HeisaFellowoftheAmer-
ican Association for Artificial Intelligence, the Association for Computing Machinery, and
theAmericanAssociation fortheAdvancement ofScience, anHonorary FellowofWadham
College, Oxford, andan AndrewCarnegie Fellow. Heheld theChaire BlaisePascal inParis
from 2012 to 2014. He has published over 300 papers on a wide range of topics in artificial
intelligence. His other books include The Use of Knowledge in Analogy and Induction, Do
theRightThing: Studies inLimitedRationality (withEricWefald), andHumanCompatible:
ArtificialIntelligence andtheProblemofControl.
Peter Norvig is currently a Director of Research at Google, Inc., and was previously the
director responsible for the core Web search algorithms. He co-taught an online AI class
thatsignedup160,000students, helpingtokickoffthecurrentroundofmassiveopenonline
classes. HewasheadoftheComputationalSciencesDivisionatNASAAmesResearchCen-
ter, overseeing research and development in artificial intelligence and robotics. He received
aB.S.inapplied mathematics fromBrownUniversity andaPh.D.incomputer science from
Berkeley. He has been a professor at the University of Southern California and a faculty
memberatBerkeleyand Stanford. HeisaFellowoftheAmerican Association forArtificial
Intelligence, the Association forComputing Machinery, theAmerican AcademyofArtsand
Sciences, andtheCalifornia AcademyofScience. Hisother booksareParadigmsofAIPro-
gramming: CaseStudiesinCommonLisp,Verbmobil: ATranslationSystemforFace-to-Face
Dialog,andIntelligent HelpSystemsforUNIX.
Thetwoauthors sharedtheinaugural AAAI/EAAIOutstanding Educatorawardin2016.
x
Contents
I Artificial Intelligence
1 Introduction 1
1.1 WhatIsAI? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 TheFoundations ofArtificialIntelligence . . . . . . . . . . . . . . . . . . 5
1.3 TheHistoryofArtificialIntelligence . . . . . . . . . . . . . . . . . . . . 17
1.4 TheStateoftheArt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.5 RisksandBenefitsofAI . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 35
2 IntelligentAgents 36
2.1 AgentsandEnvironments . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2 GoodBehavior: TheConceptofRationality . . . . . . . . . . . . . . . . 39
2.3 TheNatureofEnvironments . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4 TheStructureofAgents . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 60
II Problem-solving
3 SolvingProblemsbySearching 63
3.1 Problem-Solving Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2 ExampleProblems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3 SearchAlgorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.4 UninformedSearchStrategies . . . . . . . . . . . . . . . . . . . . . . . . 76
3.5 Informed(Heuristic)SearchStrategies . . . . . . . . . . . . . . . . . . . 84
3.6 HeuristicFunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 106
4 SearchinComplexEnvironments 110
4.1 LocalSearchandOptimizationProblems . . . . . . . . . . . . . . . . . . 110
4.2 LocalSearchinContinuous Spaces . . . . . . . . . . . . . . . . . . . . . 119
4.3 SearchwithNondeterministic Actions . . . . . . . . . . . . . . . . . . . 122
4.4 SearchinPartiallyObservableEnvironments . . . . . . . . . . . . . . . . 126
4.5 OnlineSearchAgentsandUnknownEnvironments . . . . . . . . . . . . 134
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 142
5 Adversarial SearchandGames 146
5.1 GameTheory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
5.2 OptimalDecisionsinGames . . . . . . . . . . . . . . . . . . . . . . . . 148
xi
xii Contents
5.3 HeuristicAlpha–BetaTreeSearch . . . . . . . . . . . . . . . . . . . . . 156
5.4 MonteCarloTreeSearch . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.5 StochasticGames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
5.6 PartiallyObservable Games . . . . . . . . . . . . . . . . . . . . . . . . . 168
5.7 LimitationsofGameSearchAlgorithms . . . . . . . . . . . . . . . . . . 173
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 175
6 ConstraintSatisfaction Problems 180
6.1 DefiningConstraintSatisfaction Problems . . . . . . . . . . . . . . . . . 180
6.2 Constraint Propagation: Inference inCSPs . . . . . . . . . . . . . . . . . 185
6.3 Backtracking SearchforCSPs . . . . . . . . . . . . . . . . . . . . . . . . 191
6.4 LocalSearchforCSPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.5 TheStructureofProblems . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 204
III Knowledge, reasoning, andplanning
7 LogicalAgents 208
7.1 Knowledge-Based Agents . . . . . . . . . . . . . . . . . . . . . . . . . . 209
7.2 TheWumpusWorld . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
7.3 Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
7.4 Propositional Logic: AVerySimpleLogic . . . . . . . . . . . . . . . . . 217
7.5 Propositional TheoremProving . . . . . . . . . . . . . . . . . . . . . . . 222
7.6 EffectivePropositional ModelChecking . . . . . . . . . . . . . . . . . . 232
7.7 AgentsBasedonPropositional Logic . . . . . . . . . . . . . . . . . . . . 237
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 247
8 First-OrderLogic 251
8.1 Representation Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . 251
8.2 SyntaxandSemanticsofFirst-OrderLogic . . . . . . . . . . . . . . . . . 256
8.3 UsingFirst-OrderLogic . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
8.4 KnowledgeEngineering inFirst-OrderLogic . . . . . . . . . . . . . . . . 271
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 278
9 InferenceinFirst-OrderLogic 280
9.1 Propositional vs.First-OrderInference . . . . . . . . . . . . . . . . . . . 280
9.2 UnificationandFirst-OrderInference . . . . . . . . . . . . . . . . . . . . 282
9.3 ForwardChaining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
9.4 BackwardChaining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
9.5 Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 310
Contents xiii
10 KnowledgeRepresentation 314
10.1 Ontological Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
10.2 CategoriesandObjects . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
10.3 Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
10.4 MentalObjectsandModalLogic . . . . . . . . . . . . . . . . . . . . . . 326
10.5 ReasoningSystemsforCategories . . . . . . . . . . . . . . . . . . . . . 329
10.6 ReasoningwithDefaultInformation . . . . . . . . . . . . . . . . . . . . 333
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 338
11 AutomatedPlanning 344
11.1 DefinitionofClassicalPlanning . . . . . . . . . . . . . . . . . . . . . . . 344
11.2 AlgorithmsforClassicalPlanning . . . . . . . . . . . . . . . . . . . . . . 348
11.3 HeuristicsforPlanning . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
11.4 Hierarchical Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
11.5 PlanningandActinginNondeterministic Domains . . . . . . . . . . . . . 365
11.6 Time,Schedules, andResources . . . . . . . . . . . . . . . . . . . . . . . 374
11.7 AnalysisofPlanningApproaches . . . . . . . . . . . . . . . . . . . . . . 378
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 380
IV Uncertain knowledge and reasoning
12 QuantifyingUncertainty 385
12.1 ActingunderUncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . 385
12.2 BasicProbabilityNotation . . . . . . . . . . . . . . . . . . . . . . . . . . 388
12.3 Inference UsingFullJointDistributions . . . . . . . . . . . . . . . . . . . 395
12.4 Independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
12.5 Bayes’RuleandItsUse . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
12.6 NaiveBayesModels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
12.7 TheWumpusWorldRevisited . . . . . . . . . . . . . . . . . . . . . . . . 404
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 408
13 ProbabilisticReasoning 412
13.1 Representing KnowledgeinanUncertainDomain . . . . . . . . . . . . . 412
13.2 TheSemanticsofBayesianNetworks . . . . . . . . . . . . . . . . . . . . 414
13.3 ExactInference inBayesianNetworks . . . . . . . . . . . . . . . . . . . 427
13.4 ApproximateInference forBayesianNetworks . . . . . . . . . . . . . . . 435
13.5 CausalNetworks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 454
14 ProbabilisticReasoningoverTime 461
14.1 TimeandUncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
14.2 Inference inTemporalModels . . . . . . . . . . . . . . . . . . . . . . . . 465
xiv Contents
14.3 HiddenMarkovModels . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
14.4 KalmanFilters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
14.5 DynamicBayesianNetworks . . . . . . . . . . . . . . . . . . . . . . . . 485
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 497
15 Probabilistic Programming 500
15.1 RelationalProbability Models . . . . . . . . . . . . . . . . . . . . . . . . 501
15.2 Open-UniverseProbability Models . . . . . . . . . . . . . . . . . . . . . 507
15.3 KeepingTrackofaComplexWorld . . . . . . . . . . . . . . . . . . . . . 514
15.4 ProgramsasProbabilityModels . . . . . . . . . . . . . . . . . . . . . . . 519
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 524
16 MakingSimpleDecisions 528
16.1 CombiningBeliefsandDesiresunderUncertainty . . . . . . . . . . . . . 528
16.2 TheBasisofUtilityTheory . . . . . . . . . . . . . . . . . . . . . . . . . 529
16.3 UtilityFunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532
16.4 Multiattribute UtilityFunctions . . . . . . . . . . . . . . . . . . . . . . . 540
16.5 DecisionNetworks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544
16.6 TheValueofInformation . . . . . . . . . . . . . . . . . . . . . . . . . . 547
16.7 UnknownPreferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 557
17 MakingComplexDecisions 562
17.1 Sequential DecisionProblems . . . . . . . . . . . . . . . . . . . . . . . . 562
17.2 AlgorithmsforMDPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572
17.3 BanditProblems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 581
17.4 PartiallyObservable MDPs . . . . . . . . . . . . . . . . . . . . . . . . . 588
17.5 AlgorithmsforSolvingPOMDPs . . . . . . . . . . . . . . . . . . . . . . 590
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 596
18 MultiagentDecisionMaking 599
18.1 PropertiesofMultiagent Environments . . . . . . . . . . . . . . . . . . . 599
18.2 Non-Cooperative GameTheory . . . . . . . . . . . . . . . . . . . . . . . 605
18.3 CooperativeGameTheory . . . . . . . . . . . . . . . . . . . . . . . . . . 626
18.4 MakingCollectiveDecisions . . . . . . . . . . . . . . . . . . . . . . . . 632
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
Bibliographical andHistoricalNotes . . . . . . . . . . . . . . . . . . . . . . . . 646
V MachineLearning
19 LearningfromExamples 651
19.1 FormsofLearning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651