Sunteți pe pagina 1din 5

4/14/2012

ArtificialIntelligence:Historyand Applications pp

Afterthistopic,youwillbeabletounderstand thefollowing:
Intelligenceandartificialintelligence History, History earlyworks, works andimportanceofAI FundamentalsandphilosophyofAI Modelling,simulationandAI Intelligentsystems Historyandfundamentalsofintelligentsystems

Introduction
Overthelast25years,thedisciplineofartificial intelligence(AI)hasbeengrowingonthemapof sciencesomewherebetweenpsychologyand computerscience. Researchersworkingonthetaskofmaking computersdothingswhichwouldrequire intelligenceifdonebyhumanbeings. AIisafieldofstudythatencompasses computationaltechniquesforperformingtask thatapparentlyrequireintelligencewhen performedbyhumans.

ThefundamentalissuesofAIinvolve:
Knowledgerepresentation Search Perceptionandinference

Knowledgecanbeavailableascollectionsof logicalassertions,heuristicrules,procedures, statisticalcorrelationetc.

Intelligence
HowyouinterpretAliisintelligent? Itcanbeinterpretedindifferentways:
Aliknowsalot?? Alithinks hi k fast?? f ?? Alicananswerquestionmostofthetime???

Theimportantaspectsofhumanintelligence seemtobethefollowing:
Theuseofintuition Commonsense Judgement Creativity Goaldirectedness Plausible Reasoning Knowledge Beliefs

OneappropriateinterpretationmaybeAlis actionsareappropriateinallsituations.

4/14/2012

Themeaningofintelligenceisnotthehuman brainsinformationprocessingabilitybutthe abilityofhumanstodemonstratetheir intelligencebycommunicatingeffectively, andbylearning. Communication requiresgoodskillsto analyse andsynthesismessages(todetermine thecontextandtoaltertherepresentationof information).

Learning
Theabilitytolearnoradaptonesbehaviourtoa newsituationisconsideredbymanytobeavital component p ofintelligence. g Organizingknowledgeisanimportantissueinthe learningprocess.

ArtificialIntelligence
DefinitionsofAIareresultingfromdifferent motivationsorbackground. Example,objectscanbecategorizedinto intelligent(suchasbirds birds,dogsandhuman)or unintelligentobjectssuchastable,chairetc. ThemostimportantpurposeofAIisto increasemansunderstandingofreasoning, learning,andperception.
Year 1843 1940sand 1950s Development

AI History
CharlesBabbageproposed analyticalengine,thefirstprogrammable computingmachine. AIgrows withtheemergence ofthemoderncomputer. EarlyworkbyW W.S. S McCulloch andW W.H. H Pittscanbe regardedasthefirst significantinthefieldofAI 1956 NewellandSimon solvedsimplepuzzlesandprovedtheorems in propositionalcalculus symbolicrepresentation andcomputer programming.

Earlyworks:
Workonlogic(methodsofreasoning) Alonzo Church,KurtGodel,EmilPostandAlanTuring (1920sand1930s). Cybernetics thestudyofcommunicationin humansandmachinesbecameanactiveresearch areain1940sand1950s. Late1950s GeneralProblemSolverdeveloped byNewell,ShawandSimonwrittenininformation processinglanguage.

Definitions:AIcanbedefinedasfollows:
Theautomationofactivitiesthatweassociatewithhuman thinking,activitiessuchasdecisionmaking,problem solving,learning.Bellman 1978. Theexciting gneweffort ff tomakecomputers p thinkmachine withmind,inthefullandliteralsense..Haugeland 1985. Thebranchofthecomputersciencethatisconcernedwith theautomationofintelligentbehaviourLuger& Stubblefield1993. Thestudyofhowtomakecomputersdothingsatwhich,at themoment,peoplebetterRich&Knight2003

4/14/2012

Accordingtothedefinitions,computer systemscanbeclassifiedintothefollowing categories:


Systemsthatactlikehumans TheTuringTest proposedbyAlanTuring(1950). Systemsthatthinklikehumans GeneralProblem SolverbyNewellandSimon(late1950s). Systemsthatthinkrationally concernwithlogic. Systemsthatactrationally concernwithcorrect inference.

ProgrammingMethods
Computationallogicincludestheinferenceof consequencesfromgivenfactsandrules, automatedtheoremp proving, g,logic g programming, p g g, andthelanguage. Thecomputationaluseofframe(datastructureto representinformation)forsettingupcontext. Searchingandpatternmatchingamongthebasic activitiesinAI.

Techniques
ThefirsttechniqueofAIisalgorithm. Analgorithmisaspecificsetofoperations, procedures,anddecisionswhichguaranteestoyield correctresults results. Heuristic,aruleofthumb,trick,strategy, simplification,oranyothermethodthataidsthe solutionofcomplexproblem. Therepresentationofaprobleminasimple,easily processable formthenaidsinthedevelopmentofa solution.

Fundamentalissues:
Theabilityofamachinetothink Overpromising Militaryemphasis

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

4/14/2012

IntelligentSystems
Itconcentratesonknowledgerepresentationand reasoning asthethemeonanysystemthatreflects intelligence,andsuggeststhatitispossiblefora systemtobehaveinthismanneronlyifitcontains formalmechanismsforrepresenting p gknowledge g and employinginferencetechniquesthatmodel conventionalcomputationallybasedsystems. Someimportantintelligentsystems:expertsystems, fuzzysystems,artificialneuralnetworks,genetic algorithmandevolutionaryprogrammingandswarm intelligencesystems.

Roleofintelligentsystem:
Logicalbehaviour Abilitytosolvecomplexproblems Responsiveandadaptivenature Abilitytoprovidenonlinearprogramnavigation Meanstoeffectivelyuseexistinginformation Userfriendlinessandhighinteractivity Reliability

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

ComparisonofvariousIS
Application Expertsystem Fuzzysystem Neuralnetworks Geneticalgorithms Swarmintelligence

Thecapacitytoexploitthepowersofassociation andinference inattemptingtodealwithcomplex problemsthatresembletherealworld. Theabilitytohandleanytypeofvagueness presentinaproblem. Thecapabilitytosolvecomplexoptimization problems usingsearchtechniques.

Diagnosis Optimizationproblems Prediction Controlapplications Loadforecasting Airsonartarget recognition Mobilerobotnavigation

* * * * * * * * * * * *

Tuningfuzzycontroller performance Classification ofsoil Hotextrusion ofsteel

4/14/2012

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

S-ar putea să vă placă și