SOLUTIONS & INSTRUCTOR MANUAL
Artificial Intelligence: A Modern Approach, 4th Edition
by Peter Norvig and Stuart Russell, Chapters 1 – 28
,Artificial Intelligence
U
UUUU 1 Introduction ...
U U U
UUUU 2 Intelligent Agents ...
U U U U
II Problem-solving
U
UUUU 3 Solving Problems by Searching ...
U U U U U U
UUUU 4 Search in Complex Environments ...
U U U U U U
UUUU 5 Adversarial Search and Games ...
U U U U U U
UUUU 6 Constraint Satisfaction Problems ...
U U U U U
III Knowledge, reasoning, and planning
U U U U
UUUU 7 Logical Agents ...
U U U U
UUUU 8 First-Order Logic ...
U U U U
UUUU 9 Inference in First-Order Logic ...
U U U U U
UUUU 10 Knowledge Representation ...
U U U U
UUUU 11 Automated Planning ...
U U U U
IV Uncertain knowledge and reasoning
U U U U
UUUU 12 Quantifying Uncertainty ...
U U U U
UUUU 13 Probabilistic Reasoning ...
U U U U
UUUU 14 Probabilistic Reasoning over Time ...
U U U U U U
UUUU 15 Probabilistic Programming ...
U U U U
UUUU 16 Making Simple Decisions ...
U U U U U
UUUU 17 Making Complex Decisions ...
U U U U U
UUUU 18 Multiagent Decision Making ...
U U U U U
V Machine Learning
U U
, 19 Learning from Examples ...
UUUU U U U U U
20 Learning Probabilistic Models ...
UUUU U U U U U
21 Deep Learning ...
UUUU U U U U
22 Reinforcement Learning ...
UUUU U U U U
VI Communicating, perceiving, and acting
U U U U
23 Natural Language Processing ...
UUUU U U U U U
24 Deep Learning for Natural Language Processing ...
UUUU U U U U U U U U
25 Computer Vision ...
UUUU U U U U
26 Robotics ...
UUUU U U U
VII Conclusions
U
27 Philosophy, Ethics, and Safety of AI ...
UUUU U U U U U U U U
28 The Future of AI
UUUU U U U U
, EXERCISES U U
1
INTRODUCTION
NoteUthatUforUmanyUofUtheUquestionsUinUthisUchapter,UweUgiveUreferencesUwhereUanswersUc
anUbeUfoundUratherUthanUwritingUthemUout—theUfullUanswersUwouldUbeUfarUtooUlong.
1.1 What Is AI?
U U U
ExerciseU1.1.#DEFA
DefineUinUyourUownUwords:U (a)Uintelligence,U(b)UartificialUintelligence,U(c)Uagent,U(d)Ura-
Utionality,U(e)UlogicalUreasoning.
a. DictionaryUdefinitionsUofUintelligenceUtalkUaboutU―theUcapacityUtoUacquireUandUappl
yUknowledge‖UorU―theUfacultyUofUthoughtUandUreason‖UorU―theUabilityUtoUcomprehen
dUandUprofitUfromUexperience.‖U TheseUareUallUreasonableUanswers,UbutUifUweUwantU
somethingUquantifiableUweUwouldUuseUsomethingUlikeU―theUabilityUtoUactUsuccessfull
yUacrossUaUwideUrangeUofUobjectivesUinUcomplexUenvironments.‖
b. WeUdefineUartificialUintelligenceUasUtheUstudyUandUconstructionUofUagentUprograms
UthatUperformUwellUinUaUgivenUclassUofUenvironments,UforUaUgivenUagentUarchitectur
e;UtheyUdoUtheUrightUthing.U AnUimportantUpartUofUthatUisUdealingUwithUtheUuncertain
tyUofUwhatUtheUcurrentUstateUis,UwhatUtheUoutcomeUofUpossibleUactionsUmightUbe,Uan
dUwhatUisUitUthatUweUreallyUdesire.
c. WeUdefineUanUagentUasUanUentityUthatUtakesUactionUinUresponseUtoUperceptsUfromUan
Uenvi-Uronment.
d. WeUdefineUrationalityUasUtheUpropertyUofUaUsystemUwhichUdoesUtheU―rightUthing‖Ug
ivenUwhatUitUknows.U SeeUSectionU2.2UforUaUmoreUcompleteUdiscussion.U TheUbasicUc
onceptUisUperfectUrationality;USectionU??UdescribesUtheUimpossibilityUofUachievingUpe
rfectUrational-UityUandUproposesUanUalternativeUdefinition.
e. WeUdefineUlogicalUreasoningUasUtheUaUprocessUofUderivingUnewUsentencesUfromUold,U
suchUthatUtheUnewUsentencesUareUnecessarilyUtrueUifUtheUoldUonesUareUtrue.U(NoticeUth
atUdoesUnotUreferUtoUanyUspecificUsyntaxUorUformalUlanguage,UbutUitUdoesUrequireUaUw
ell-definedUnotionUofUtruth.)
ExerciseU1.1.#TURI
ReadUTuring‘sUoriginalUpaperUonUAIU(Turing,U1950).U InUtheUpaper,UheUdiscussesUseveral
UobjectionsUtoUhisUproposedUenterpriseUandUhisUtestUforUintelligence.UWhichUobjectionsUstillUcar
ry
©U2023UPearsonUEducation,UHoboken,UNJ.UAllUrightsUreser
ved.