You are on page 1of 19

1. What do you meanbyintelligentagents.

ExplainSimplereflex,Goalbased,modelbased,
utility based and learning agents. State the limitations of each of them and how to
overcomeinothertypeofagent.

A)ReferIST1

2. Describe the Hill Climbing Search Algorithm. What are the problems faced by Hill
ClimbingSearch?Suggestthemethodforeachproblemtoovercome.
A)Refer3.5.3.2tillpg351abovetheBlockWordProblemtillthesolutions.

3. How to solve the problem of deciding how and when to evaluate the intelligent agents
success?Distinguishbetweenrationalandautonomousagent.

A)Refer2.2

4. WhatisAI?ExplainthedifferentcomponentsofAIwithablockdiagram
A)AnswerSearchUnderProgress

5. HowBestFirstSearchwillbeapplicableto8puzzlegame?
A) Write the algo of Best First Search andthendirectly3.5.3.3pg358(afterdrawback)tillnext
pgaboveandexcludingroutefindingproblem

6. Give the former grammar for subset of English and designandgivethepseudocodefor
Englishtutoractingasexpertsystem.Statetheassumptionsclearlyifany.
A)Referthe11RsXeroxDec2004Q1)a

7. ProveA*algorithmiscompleteandoptimalamongallsearchalgorithms
A)Refer3.5.3.4uptopg363tillModificationstoA*

8. GivetheDepthfirstsearchalgorithmandstateitsadvantagesanddisadvantages
A)ReferIST2

9. WhatisanExpertSystem.Explainitalongwiththearchitectureandcharacteristics.
A)ReferIST1

10. Considerthefollowingstatements:
Ifthemaidstolethejewellery,thenthebutlerwasnotguilty
Eitherthemaidstolethejewellery,orshemilkedthecow
Ifthemaidmilkedthecow,thenthebutlergotthecream
Therefore,ifthebutlerwasguilty,thenhegothiscream
Questions:
ExpressinPropositionalCalculus.
Expressthenegationinclausalform.
Demonstratethattheconclusionisvalid,usingresolutioninpropositional
calculus.

A)ReferIST2

11. Whatdoyoumeanbyheuristicsearch?Illustrateitwithhillclimbingalgorithm.
A)ReferIST1

12. Suppose thatadatawarehouseconsistsofthefourdimensionsdate,spectator,location,
and game, and the two measures count and charge, where charge is the fare that a
spectator pays when watching a game on a given date. Spectators may be students,
adults,orseniors,witheachcategoryhavingitsownchargerate.
i) Drawastarschemadiagramforthedatawarehouse.

ii) Starting with the base cuboid [date, spectator, location, game], what specific
OLAP operations should you perform in order to list the total charge paid by
student,adultspectatorsatMumbai,PunePlacein2013,2014?

starschema
A)


13. What is a training set andhowisitusedtotrainneuralnetworks?Belowisadiagram


ofasingleartificialneuron(unit):


The node has three inputs x =(x1,x2,x3)that receiveonlybinarysignals(either0
or 1). How many different input patterns this node can receive? What if the node
had four inputs? Five? Can you give a formula that computes the number of
binaryinputpatternsforagivennumberofinputs?

rainingset
A)T

Thefirstsubsetisthetrainingset,whichisusedforcomputingthegradientandupdatingthe
networkweightsandbiases. Trainingsetcanbemadeeasilydirectlyfromthetimeseries.
Certainnumberofmeasuredvaluesisusedasinputsandthevaluetobepredicted(i.e.,the
valueinthefuture,insomechosendistanceaftertheseinputmeasuredvalues)isusedas
requiredoutput.Inputpartofthetimeseriesiscalledwindow,theoutputpartisthepredicted
value.

TraininganArtificialNeuralNetwork
Inthetrainingphase,thecorrectclassforeachrecordisknown(thisistermedsupervised
training),andtheoutputnodescanthereforebeassigned"correct"values"1"forthenode
correspondingtothecorrectclass,and"0"fortheothers.(Inpracticeithasbeenfoundbetter
tousevaluesof0.9and0.1,respectively.)Itisthuspossibletocomparethenetwork's
calculatedvaluesfortheoutputnodestothese"correct"values,andcalculateanerrortermfor
eachnode(the"Delta"rule).Theseerrortermsarethenusedtoadjusttheweightsinthe
hiddenlayerssothat,hopefully,thenexttimearoundtheoutputvalueswillbeclosertothe
"correct"values.

Aneuroninanartificialneuralnetworkis

1. Asetofinputvalues(xi)andassociatedweights(wi)
2. Afunction(g)thatsumstheweightsandmapstheresultstoanoutput(y).


14. Describethemainstepsofthesupervisedtrainingalgorithm?

A)Supervisedlearningisthemachinelearningtaskofinferringafunctionfromsupervised
trainingdata.Thetrainingdataconsistofasetoftrainingexamples.Insupervisedlearning,
eachexampleisapairconsistingofaninputobject(typicallyavector)andadesiredoutput
value(alsocalledthesupervisorysignal).Asupervisedlearningalgorithmanalyzesthetraining
dataandproducesaninferredfunction,whichiscalledaclassifier(iftheoutputisdiscrete,see
classification)oraregressionfunction(iftheoutputiscontinuous,seeregression).Theinferred
functionshouldpredictthecorrectoutputvalueforanyvalidinputobject.Thisrequiresthe
learningalgorithmtogeneralizefromthetrainingdatatounseensituationsina"reasonable"
wayasupervisedlearningalgorithmtakesaknownsetofinputdataandknownresponsesto
thedata(output),andtrains
amodeltogeneratereasonablepredictionsfortheresponsetonew
data.

Forexample,supposeyouwanttopredictwhethersomeonewillhaveaheartattackwithina
year.Youhaveasetofdataonpreviouspatients,includingage,weight,height,bloodpressure,
etc.Youknowwhetherthepreviouspatientshadheartattackswithinayearoftheir
measurements.So,theproblemiscombiningalltheexistingdataintoamodelthatcanpredict
whetheranewpersonwillhaveaheartattackwithinayear.

Supervisedlearningsplitsintotwobroadcategories:classificationandregression.

Inclassification,thegoalistoassignaclass(orlabel)fromafinitesetofclassestoan
observation.Thatis,responsesarecategoricalvariables.Applicationsincludespamfilters,
advertisementrecommendationsystems,andimageandspeechrecognition.Inregression,the
goalistopredictacontinuousmeasurementforanobservation.Thatis,theresponses
variablesarerealnumbers.Applicationsincludeforecastingstockprices,energyconsumption,
ordiseaseincidence.

Inordertosolveagivenproblemofsupervisedlearning,onehastoperformthefollowingsteps:

1. Determinethetypeoftrainingexamples.Beforedoinganythingelse,theengineer
shoulddecidewhatkindofdataistobeusedasanexample.Forinstance,thismightbea
singlehandwrittencharacter,anentirehandwrittenword,oranentirelineofhandwriting.
2. Gatheratrainingset.Thetrainingsetneedstoberepresentativeoftherealworlduseof
thefunction.Thus,asetofinputobjectsisgatheredandcorrespondingoutputsarealso
gathered,eitherfromhumanexpertsorfrommeasurements.
3. Determinetheinputfeaturerepresentationofthelearnedfunction.Theaccuracyofthe
learnedfunctiondependsstronglyonhowtheinputobjectisrepresented.Typically,theinput
objectistransformedintoafeaturevector,whichcontainsanumberoffeaturesthatare
descriptiveoftheobject.Thenumberoffeaturesshouldnotbetoolarge,becauseofthecurse
ofdimensionalitybutshouldcontainenoughinformationtoaccuratelypredicttheoutput.
4. Determinethestructureofthelearnedfunctionandcorrespondinglearningalgorithm.
Forexample,theengineermaychoosetousesupportvectormachinesordecisiontrees.
5. Completethedesign.Runthelearningalgorithmonthegatheredtrainingset.Some
supervisedlearningalgorithmsrequiretheusertodeterminecertaincontrolparameters.These
parametersmaybeadjustedbyoptimizingperformanceonasubset(calledavalidationset)of
thetrainingset,orviacrossvalidation.
6. Evaluatetheaccuracyofthelearnedfunction.Afterparameteradjustmentandlearning,
theperformanceoftheresultingfunctionshouldbemeasuredonatestsetthatisseparatefrom
thetrainingset.

Awiderangeofsupervisedlearningalgorithmsisavailable,eachwithitsstrengthsand
weaknesses.Thereisnosinglelearningalgorithmthatworksbestonallsupervisedlearning
problems

15. Supposeageneticalgorithmuseschromosomesoftheformx=abcdefghwithafixed
lengthofeightgenes.Eachgenecanbeanydigitbetween0and9.Letthefitnessof
individualxbecalculatedas:
f(x) = (a+b)(c+d)+(e+f)(g+h), andlettheinitialpopulationconsistoffour
individualswiththefollowingchromosomes:


i) Evaluate the fitness of each individual, showing all your workings, and arrange
theminorderwiththefittestfirstandtheleastfitlast.
ii)Performthefollowingcrossoveroperations:
Crossthefittesttwoindividualsusingonepointcrossoveratthemiddlepoint.
ii) Cross the second and third fittest individuals using a twopoint crossover
(pointsbandf).

geneticalgo
A)


16. Solve Using kmedoid algorithm Cluster the following data set of ten objects into three
clusters i.e. k = 3. Calculate cost using Manhattan distance. Consider a data set of 8
objectsasfollows:
A1(2,10),A2(2,5),A3(8,4),B1(5,8),B2(7,5),B3(6,4),C1(1,2),C2(4,9).
Thethreeclustercentersafterthetworoundofexecution.

ANSWERSEARCHUNDERPROGRESS

17. Consider the followingdatabasewithminimumsupportcount=2.Findallfrequentitem


set using Apriori and also generate strong association rules if minimum
confidence=70%.

A)Apriori

18. Describethemainstepsofthesupervisedtrainingalgorithm?

A)SAMEASQ14

19. Describetheworkingprincipleofgeneticalgorithm.

ReferQ4)bofMAY2005ofthe11RsXerox

You might also like