Sei sulla pagina 1di 5

4/14/2012

Afterthistopic,youwillbeabletounderstand
thefollowing:
ArtificialIntelligence:Historyand
Intelligenceandartificialintelligence
Applications
pp History,earlyworks,andimportanceofAI
History early works and importance of AI
FundamentalsandphilosophyofAI
Modelling,simulationandAI
Intelligentsystems
Historyandfundamentalsofintelligentsystems

Introduction
Overthelast25years,thedisciplineofartificial ThefundamentalissuesofAIinvolve:
intelligence(AI)hasbeengrowingonthemapof
sciencesomewherebetweenpsychologyand Knowledgerepresentation
computerscience. Search
Researchersworkingonthetaskofmaking
Researchers working on the task of making Perceptionandinference
Perception and inference
computersdothingswhichwouldrequire
intelligenceifdonebyhumanbeings. Knowledgecanbeavailableascollectionsof
AIisafieldofstudythatencompasses logicalassertions,heuristicrules,procedures,
computationaltechniquesforperformingtask statisticalcorrelationetc.
thatapparentlyrequireintelligencewhen
performedbyhumans.

Intelligence
HowyouinterpretAliisintelligent? Theimportantaspectsofhumanintelligence
seemtobethefollowing:
Itcanbeinterpretedindifferentways: Theuseofintuition
Aliknowsalot?? Commonsense
Alithinksfast??
Ali hi k f ?? Judgement
Creativity
Alicananswerquestionmostofthetime??? Goaldirectedness
OneappropriateinterpretationmaybeAlis Plausible
actionsareappropriateinallsituations. Reasoning
Knowledge
Beliefs

1
4/14/2012

Themeaningofintelligenceisnotthehuman Learning
brainsinformationprocessingabilitybutthe Theabilitytolearnoradaptonesbehaviourtoa
abilityofhumanstodemonstratetheir newsituationisconsideredbymanytobeavital
intelligencebycommunicatingeffectively, componentofintelligence.
p g
andbylearning. Organizingknowledgeisanimportantissueinthe
Communication requiresgoodskillsto learningprocess.
analyse andsynthesismessages(todetermine
thecontextandtoaltertherepresentationof
information).

ArtificialIntelligence AI History
Year Development
DefinitionsofAIareresultingfromdifferent 1843 CharlesBabbageproposedanalyticalengine,thefirstprogrammable
motivationsorbackground. computingmachine.
1940sand AIgrows withtheemergenceofthemoderncomputer.
Example,objectscanbecategorizedinto 1950s
intelligent (such as birds dogs and human) or
intelligent(suchasbirds,dogsandhuman)or Early work by W S McCulloch and W H Pitts can be regardedasthefirst
EarlyworkbyW.S.McCullochandW.H.Pittscanbe
significantinthefieldofAI
regarded as the first

unintelligentobjectssuchastable,chairetc. 1956 NewellandSimon solvedsimplepuzzlesandprovedtheoremsin


propositionalcalculus symbolicrepresentationandcomputer
ThemostimportantpurposeofAIisto programming.
increasemansunderstandingofreasoning,
learning,andperception.

Earlyworks: Definitions:AIcanbedefinedasfollows:
Workonlogic(methodsofreasoning) Alonzo Theautomationofactivitiesthatweassociatewithhuman
thinking,activitiessuchasdecisionmaking,problem
Church,KurtGodel,EmilPostandAlanTuring solving,learning.Bellman1978.
(1920sand1930s). Theexcitingnewefforttomakecomputersthinkmachine
g ff p
Cybernetics thestudyofcommunicationin withmind,inthefullandliteralsense..Haugeland 1985.
humansandmachinesbecameanactiveresearch Thebranchofthecomputersciencethatisconcernedwith
areain1940sand1950s. theautomationofintelligentbehaviourLuger&
Late1950s GeneralProblemSolverdeveloped Stubblefield1993.
byNewell,ShawandSimonwrittenininformation Thestudyofhowtomakecomputersdothingsatwhich,at
themoment,peoplebetterRich&Knight2003
processinglanguage.

2
4/14/2012

Accordingtothedefinitions,computer ProgrammingMethods
systemscanbeclassifiedintothefollowing Computationallogicincludestheinferenceof
categories: consequencesfromgivenfactsandrules,
Systemsthatactlikehumans TheTuringTest automatedtheoremproving,logicprogramming,
p g, g p g g,
proposedbyAlanTuring(1950). andthelanguage.
Systemsthatthinklikehumans GeneralProblem Thecomputationaluseofframe(datastructureto
SolverbyNewellandSimon(late1950s).
representinformation)forsettingupcontext.
Systemsthatthinkrationally concernwithlogic.
Searchingandpatternmatchingamongthebasic
Systemsthatactrationally concernwithcorrect
activitiesinAI.
inference.

Techniques Fundamentalissues:
ThefirsttechniqueofAIisalgorithm. Theabilityofamachinetothink
Analgorithmisaspecificsetofoperations,
procedures,anddecisionswhichguaranteestoyield Overpromising
correct results
correctresults. Militaryemphasis
Military emphasis
Heuristic,aruleofthumb,trick,strategy,
simplification,oranyothermethodthataidsthe
solutionofcomplexproblem.
Therepresentationofaprobleminasimple,easily
processable formthenaidsinthedevelopmentofa
solution.

CurrentTrendsinAppliedAI
Thekeytechniquespresentlyinuseareexpert Intermsofpracticalapplications,duetotheir
systems,fuzzysystems,neuralnetworks, learningcapabilities,neuralnetworksare
geneticalgorithmandswarmintelligence. oftenemployedwhereaproblemrequires
p y y
AnexpertsystemdifferfromotherAIsystems estimation,prediction,orclassification.
inthatitattemptstoexplicitlyembody GeneticAlgorithmsarefavoured forsearch
expertiseandknowledgewithinsoftware. problemswhichrequiretheidentificationofa
Ausefulapplicationofsuchfuzzysystemsisin globaloptimalsolution.
decisionmakingunderuncertainor Swarmintelligenceutilizesthebehaviour of
hypotheticalsituations. nonhumanlivingentitiesforproblemsolving.

3
4/14/2012

IntelligentSystems
Itconcentratesonknowledgerepresentationand Roleofintelligentsystem:
reasoning asthethemeonanysystemthatreflects
intelligence,andsuggeststhatitispossiblefora Logicalbehaviour
systemtobehaveinthismanneronlyifitcontains Abilitytosolvecomplexproblems
formalmechanismsforrepresentingknowledgeand
p g g Responsiveandadaptivenature
Responsive and adaptive nature
employinginferencetechniquesthatmodel
conventionalcomputationallybasedsystems. Abilitytoprovidenonlinearprogramnavigation
Someimportantintelligentsystems:expertsystems, Meanstoeffectivelyuseexistinginformation
fuzzysystems,artificialneuralnetworks,genetic Userfriendlinessandhighinteractivity
algorithmandevolutionaryprogrammingandswarm
intelligencesystems. Reliability

Comparisonwithconventional
programs
Traditionalcomputerapplicationssuchas Theydrawknowledge,andhavethepowerof
databaseandgraphicsprogramsarebasically associationandinference,enablingthemto
timesavingtoolsthathavereplacedmanual steerthedirectionofarunningprogram
approaches totaskssuchascalculating,
to tasks such as calculating towards useful results
towardsusefulresults.
sortingandsoon.Theseprogramssimply Intelligentsystemspossessthefollowing
employalgorithm andexhibitnointelligence. importantproperties:
Intelligentsystemsaremoreflexible and Theabilitytouseknowledgetoperformcertain
adaptive. tasksorsolveproblems.

ComparisonofvariousIS
Application Expertsystem Fuzzysystem Neuralnetworks Geneticalgorithms Swarmintelligence

Thecapacitytoexploitthepowersofassociation Diagnosis *
Optimizationproblems * *
andinference inattemptingtodealwithcomplex
Prediction *
problemsthatresembletherealworld. Controlapplications *
Theabilitytohandleanytypeofvagueness Loadforecasting *
presentinaproblem. Airsonartarget
recognition
*

Thecapabilitytosolvecomplexoptimization Mobilerobotnavigation *

problems usingsearchtechniques. Tuningfuzzycontroller * *


performance
Classification ofsoil *
Hotextrusion ofsteel *

4
4/14/2012

Reference
N.P.Padhy,2006,ArtificialIntelligenceand
IntelligentSystems,OxfordUniversityPress,
NewDelhi,India.

Potrebbero piacerti anche