UNIT -1 Se ee a me et me me i lO he Aah an ke a em| Aersriciat Laitert4ence (AT) Astificial_ intelligence _to. make Computers do sings a _ak shi | tne moment, people Bae cin NaDidal inkelliqence is Mal Guanch of Computer Science “dealing with Symboli 3 il at | Ie Psocessing _ond mone lgoaileSal mebnods off Qnoblewn Solving” a = | WistoRY = es a i ae fe saat Bian re Sar Stat of "ag Sea i |] Computes. oe AL ca it fo Compyates: i [ lec i Fi Was 3964 -4ay0_sAa2 SABE a0 ay Jaana 24Ka A956 AAS% [ Lage lenges’ — meron — Sto sequen Tity Commerial bog Reema ga ee eee hag toe ey vt ‘3 ‘Aevealed C hessmasien 7 ¥ With tne development of tne ¢leclxonic_ Computer Bn sau, | Ane kechno log finally \ecame jlable to Create wnachine_ — Inte igence in 1444, Ine Stosed pioquamn Computer inate — me jor of Entering a ¢ ew aa) alugnresn ents putes near Compsker Science ond intelligence “ine form Maficial eventvally, psnpicial vie tas fist Coined jn 4456 and then— | tnrettigence Aunficiad Intelligence hha L£xpanded tn 1458 Me Ga ation mew de opment the Lise lang wage announced Myusp stands for List ao deceived a 22 + million. ¢ x Jes Lore $e cut | ! us govedument 40 be A el iv ing Maching——J oy ogni ribion, Dung me 12% iS in tne Aevelnoment op a were. ate for Cxommple —__| Vow it vould re po! sible lo Al Ving vislh_on0. ienasye | 4 ase’ on Me shadk 4, of oP image Wasic infosmabion | & on s, Coloe ellges A fede Sn 86 ys BI is soles of AL-Aclated ae and Sfuy to $'aSmiNim | 16 mas in pa akiculad demand because ale op Yneia Efficiency. Comgantes. Quch_as Dias skal t | ielecdannt woe using an fxperk Sug Fern iS ane, | t to Pngdam tne lege Canton Buk Now in 196 Tas stout Ree Uet ON ee techno logs et Bon ag Fv22¥ \ogic Dga_nevsal nek works were i being recansi escd a possible Wows ep _achievisoty + akpiciol \ntel\isen ce ld | APs She EtonS 5 ol ji { aT EXPERT System 4 i! Expoat Su Jerns ane AL: ero qdorn: Ynat maniac : cod | [| FI domain on area of Expea hice. “Tnese asco Aesigned 40 _—] | i [assists expats not neplace them. “The oxea! of apelicati a \e i | a garlieu! human expertise is dequised in Sdve_pactlems ,— | export System Can Ise Wek : icaors gens: of | expest mre Soc as cA Nernicet | nels s Ganking Cornputes Syskern Con hay welisreee : i | , | . | | ‘The Net gees: onvand on Some Caacaples Of Expert System : —ithak ate in ise ae = MYCIN = in Medical Epier aes isa 5 |-— Pros pe. Sealogy ana.) in Computreas.. e = beh ne i bide 2) ftusat. Lawavace mats ROCESSING a vt Notun 106 2 of oui Soutine ay go in_uahich Le «ais ee 1 human $0. Sil fel ~ end wee: sti age tying Rica CRS SQ > much tatalligenk Prabal indessiand tne natusal Jeangus a6 Xx Such _as ai h_08, any othes languages. u 3 Nip iid Age eRe aa ~ya_| Natusal eguagy_Undettaating = so ind nella hing 2 > ingtiuchion gen ba i Yak | x ~¥ Onderstand eaple moze: eagl?, Bs Rerum corn See 1 nde < 2. Genetalion of natosal language Sa tat pengle Carr ~. qj—Understand Computers ‘mare Ea, ily Ee 2 Speoch secognitian © Spoken language isa more naluiat | 2 Gorn of Communication tn many homen Cow, nh $0 bral Ane Com heag out voice md bo 8 in nese are § maion design issues in Soccch Sy tern Zz 8 dependence V Kot inde el 2 peaker Independent System can Ii en m4 Secakes = Nate Sound ho to: dependence 5 Systern understands MWe paltoms af & SpeakeR i] z PE —Continvers Vs icalated word Speech = tu isola¥, - Wwoid seh sme. Speaker Pauses blur fach Word Inete cad a 4 Con Hnveus, Speech ai aah—_ @ Realtime Versus Offline Processing, ¥———— Highly, \nkoaachue application equine back a Sentence | te Aransiaked inten berks Ih is being Spoken, while —— Tin einer Gibrolten sik is pei asible— so Spend ————_| Anukes In Comapubalion dius |nemov quosares | qi 4 BK eee here K 1s any nombea tele Zax | Games PLAYING = To tim 1903, ne Weosld numer one playea-of sss a Le Garay was defer kod bu 1gn's Dee: BQ Winn ©. I Newsweek Magazine descailred le ina “sme acin's lame hand” wong ko ve Vv Making Computers play games See: Inkeressin 4 nels 1s tay mans fox Comeuker game Playing me: archers Mav: ws a a machine: fo ela l_ Vic -kac-loe. tn ne NA ko. jest oeerationa 0 _ jo Sorne playing AAS vitue ales aie limiked Less Knowledge ts Aequized. od | dame playina—paogiare to clay chinese checker, ———| an a. . of ws: |_ “Thee ase following Seasons. Es +| Games provide @ Shavclusal bask tr _talnich. HAL is vexy Gas to measure Ss os fatluze: -| Game playine, Pests. eto Stinula te t | reel life SihisaHens.» w_| ot CompuTeR vision This Simple task to attach a Camera to o Computer — So Hate Computer can seceive viguad images. tb has 40 be a wore AtPficulk task to. thon pred dose __ images So Yhak he Computes Cam understand i aes Whak_ik is Seeing __We genesall 4 See more —Yhan nea Pe Aine_Jeal of Computaa_ neem ane " is Same pomresful facilthy faa __. __ bvderskanding SCA here age age ey ‘ = 1 MpPiewe bask \ f Pies = ee age is tano- Aimensional, while te vail i 3 nice dimensional . Same iagaeree races = (shen cm 4s age in Created ___=____ 4 2 ee Some} > olsjerk nay Pas tia\ Jade abner. ee 4 we J 7 o ‘\ Agi > Tne _ualue of ingle “pixe\ is. affected. on. 5 a0. gy dn, 1 eck, dene if ova +, and _dislane of (amaxa ele. a Anal as weit mecwnan! i iL Con le pangdammed to pexfarm manual +e STL AN dototies § nok cask of AL.” A sobok perf Peal Yose Punchons Por wich {bjs Sroqsammed bub 2 | the Inreliigent robot indodes gome Kind of Seasoas ™ fi) Sch as Comua tngk allows iL bo mespand $o chamges in =| | its favuonmenk Sabnex Han just jo allow in shiechor $fel e Grek asne wat cOh Ree enn aes Psat oh Pexfoams ine ob tat tat — quate se Taig Bo ik] HS —t ~ sas = ak Gives de oe aet eae! ie ‘yakious ae A aaa wipulaiel. atads_at-an_tniegace —[—f- jp __ji Ek Lnulionmont ig 4 +1 Powsoe Susply nil sb ginuides necessaiy entity kn | ' pap BD BB h& hi hfe & @ @ | _Findima a Pana for ao thea Sem in mathematics | can we" aeqarded asa Stites b, \_ ask. Not | only _doe: WW wequire ne alnliky lo make Seductions (ram yp neces”, also Also nich hel kx) One na ernalicia: a hag been Siqnifi ntelliqence meted “ne furalnaton dedvelive piace tne age of medricate oy elos VS lo under stan ets Ae? of ne Components of Seasoning | Many lak eluding rmedical } Infoamation pebiev. cen tre Poavndlined as then tern gacsind, i? eee Ree KisAtG P. KNGWLEY oea Sie » KNOWLEDGE $ KNOWLEDGE Preamp * %__» Kmovled is defined as Facts , 2 in fo&mation and Skles acquired through » experience on edu cation 9 ES ane Nierench of know Sedge Pyramid i ECL ENN Dot ia SMown in 4'4- elow >. > ’ 5 InteLugence 2 Knowtepse 4 IweoR MATION 3 : ak Symgor S } ie 9 4+ SymevLs + Ab the base of He knowl edge y Pusamid ane Symbols which form tne ; means ot nepresentaton. , ' Sy Os) Danes, a. INFoRmATION = AnfoAmation is tne Processed dala on Which cecisions , and acHong are based. ’ , ue KniowLe 0oe £ On-foxmahon tons-fosmed , inlo meaning: kines uted is Oganized infoama tion prot hePes in decision ha:6 conve UG EN Pee Abas a spo 2068 one ae 4 Process - a+ is the ability ul usePul infekences from koe BUS Knowledge: hit 6 WisDom = foacomos tne of ne rind Fat Monee Ss 1 2 Mi exrre to achieve desirable see € « « « € € € ‘ ¢ « s se eg LS in ie Rig aPEOPLE 7 CompuTEs = WuAt Computers CA DO BeiteR Tan “PEOAE ROM ONES ee ee ramen ea, eee > Thee ane won oe that Computers Cay do better Wan People : nlumesicol Computation © one Small, SpecioD = purpose Com puler voith amilion is tne Cal cwlalor- vohich ae ane Ver’ ees? oS wene askes +0 Divide] mod Holy an rad a: pines wy 68. = wovdd Yor penfoarn tne Computaton im your : Mead One Usene Calcvlator’ » which technique wovdd be fasten. sp Jj amone Aaccvnake? ps Gornputens one So fast 4nat Yney Can p Perfoam calculations that ltkenally loold > be im possi ble fom one person Eo Compl ete » ma SieHrne. Fees ; = ee Tnfoamato storage 5 might Contain, for » > ~ business Computer y Cxample , the mame addnesses thovsends : Cusgtomeas , incledin b Cornplete guecoads dheide financial trangachons. ac xf yoo didv't hove a Computer, do you $ oppose inatesiee ould) beepessible foe eqov +o gnemember ei of that datay es S00 Compv ter ‘Con Nremewabea mach amount of information and srecall anny of ik ak “Yous Command. |3 Repetitive Openattons + > tinedness f bomedom ane the chanac benistes of man nob 9, machine. Machine Can do tne Srepe titive jos Withovt geting boned 3 fined. We eqvent instaoct Compuers to Perform dhe Same taske again § agains for example , +o pamt 000 Copies a Cvstomr ne pont Bot If we do this job ynanvabhy , downfall in “the quatity Loigk be ob coaved toith: the passage of Lithe dve to boxedorn and fabigue- & VERSATILITY & A Computer 1s a Versatile machine, This machine Con be Used to Solve tne problerns Aelaling +o Vaniovg different fields. Compuken is mulk tasking System i-e- Voriovg task pernfoAm at the Same time but @ Ppersom Can not do Various task at Same time. pa Ken SVEN tne Computer is a weliable elec taonic mull puapose and mull Psoces-Sin machine. Tt processes data wilh high aceunac y witnout any mistakes: & DILIGENCE F A Computes Pia Opendt eee oo Continous] Luttnouk faking any west: I+ User wim Sore Concentration Pita be eailian sn otrochonS ane to be penfonmed ak oiht perform ne ingrhuchon with xe Same accuracy lt} yi dtion nm ung ne finsk instauchon.) ) 2 ge WHAT PEceLE CAN Do BeTreR Than Computer + : s2People have tnaditionally ovtperformed , Computers hn achviltes that fnvolve fntetliqence This Veny dippicvlk to define tne ona d ) mteiliqence. But we Can Say that nM the , allowing ackvihes inyellig ence is Shown by ¢he® humen beings. we do much mone Than just Processin dye infoAmahon, We Undeastand it. |e We “make Sense” Ovt of what we See 3 | head and Come vp wilh nero ideas . » we use Common Sense to make dur va thaough the iWogical Problem s PR KE! Gocill AT is to design te Syskem to panfobve ine Gelvines peels inteMiqence becavse we believe that ™ oth Hie beter than machine. acivibies a mofo us 1644 ‘So me goal pr ts to make Cormpolus mone inte Iigen 2% Unstauchuaed Problem - Solving > Soin oA Problems In which te gules do Not comently etst- EArt a docto diag-~mosin a disease, & Nowy ax losing aw Persuasive Awgunen\ ele. using Common Sense to Solve slew Pao blems > gk is ovr abiBiky go “look at okjeckS we have mover seen wafine and apely ovr Comm ov) Sense to Undershemd how they wotlk fk sna Wwe need fo do, in me Sitvalion ica i cFecung Emotions § Uuset Sia AME. Emotion$ OF OTHERS & S —> “There we gobots that Can precog M3 mow, dike Pepper , a Socal wacker erat Co" Jand Cmotlons P Speak wilh eeople e €mokons Unders => Pepper hag been PAD npammed to grecpond, bok it doesn't fr\\y Gndaskend Lohat vumans one feeling - Homans fee\ em path 4 making them beler Cane Lakers - 5. QREATIMTY © intelli COO gang Con — Aak ficially Create beavk uf Nusion of photos 0% mimic a famovs orkis4's Style. Qutb making Qa work of ont people wilt remnady & uniquely human Skit — Tris possivie to design an Con moke on endless Sequence bot ik is Ufo lo beach an algoat¥onn we beeen We moh onally Provoarf oR adgoaitnn dat paintings. 7 a aPuence and me Aneck. _ ce ey 2, eco do not generally, Kespond tne Lame 4 Zach Hme while talking witn aM tdenheal people. aePROBLEM CnagacTERISTIES + > In Ondex to Choose Me wost appropriate menod for Oo poakiculak problem, it is NeCceSsanr to check tne Problem alone, different Key dimensions 4s DecompoSasre To SmarreR PRowrems + ——> 1s Mme problem decorn posable into a Se of independent Srnaller or Easier Sob Problems. a: PREQICTASLE PROBLEM YuweERSE F ne prodlen universe is predictable, Wwe fon Lo enensle a Sequence opecations é a te —— make a elon nak ts quosanteed Jo lead to % Sorvtion StePrs Cpr Be lquoreD OR Unoone — a step ts Wtong , WE Can backhraclk and asremer O diferent approach - Tf we ake prying oe Jame cress amd nearize & ~mishke offer & Couple of moves: We Cannot Simply ‘enore tne aniskake and packkrack. 4 Good Solutions ARF OBviIoVS > —— \s a dood: SoluHon to me PAobsem obvious ss\tno ot Comparison 30 al! omer possible SolvHon? 5 REQuiRES LoTS OF KNOWLEDGE OR VSES knowlensge To ConsTeAin Solutions + ——> \s a lage omount of K Solve we Prodlemn ot is Kno to Consticin tne Search? Be Solumion) A State, OR Al FATHTOSesSTANe — > |s_ ne desixed Solotton & State oF O Pah Eo Oo Stare mow\edge nequixed 40 wed ye fmporiant onlyfe Do eS ASE REQVIRE HUMAN — InreER ACTION = ——> Compu tn Simply nekan Me Solvion of yne qiven paoblem of Lilt We SoloHon tne problem neque interacHon bf te Computer. and a pearson: ie an Example of =F 7et8 Problem Choaaciais Hes aca Below . < Water Su >a oblem being applied to Solve & ey e Problem Crorackerishe Saristied| ‘Reason Ig dhe Problem decompsablet No One Single Selokion Con Salvvon Steps be ignored yes on Undone? Z 1s ne Problem Universe YES Problem Universe is Predickable| Predictable? ber fo Solve Mnig prodlern tt 2 equire only one Paxson TS a Qood Solovion absolule | Absolute | Absolute SolvHon, Water OH Kelakve 7 Jog Problem may have number of Sclotion ) bot once we ‘found one Sot mo need to boner aboot Omer ScloHon- 1s a SoluHon a stale oF O | Pa IH Patn alousolavlen Pow 7 thot is tne vole of Lot of Knocaledge hetes Knowledge ? SLC ee ee for a Solution Does me tsk Nequire Yes AddNona® assis tare is Noman - \nkeaacton Kequized Mee to gee Jugs oa pumeProplem Representation In AL Problem Con be Sie presonted by follousin amemo 43 = 4 4. Stare Seance Representation 2. Proatem ReBuction ae Grave Sence RepreSen TATION F i med in ne = In eis snernod; Problem, is dof {orn Zlakes. Here gjake means tHe Position Tar yak Awe - ; 5 When we define Problems, Accoadin to Slates, Problem become” easy and mone Undarstan da ble es A give Space = Aepaesented by a fa0r tulle os Koller > 4 Nis ene Se Peco Sa is me Sek of Aace on Banks ble nodes Re Ye mon emery Subset a NI, Contains me Siar Stare of ine Problem: 4. GH,a non empty gvubset et nN, Conteins me oat Siate af re problem » Mowe to make & cue of lee Example + : Sa anig Ine {nak Stake is Suveatle bea. Bolling, a> Steating slate Is: tne dake Adding Ten | Waren Water. ~F Collec ® Fodder Adding MILK IngAe dientS accaading to nedsiszemen’S PSLodan. — 5 Add all things in @ Seq venre Accoading so poanstn4 a CuO of ‘¢ea- TEA ern apenas aco [SeNEneLe]rs Problems in gree Space Representation or possible to sae abe stares for & given puoblews ae TS Bat entienel Cone aan 2 PROBLEM REOY CTION! + ROG CES ae _ 5 To ovnccomie. ne Problems et siete Spore xnetned ; Guede tN technigue is Sub divided Into Sede pieces: lex Problern if ik Problem noduc How Complex problem is ae 3S Jogian to Solve & Come ig divided into Sub problems aot Zaamp\e 4 we have to actxornge a dinn {x %-lo puson ot ome we cam aei¥ easily bok 1f we have to a4Aan a Ainner {or 20° People ih is difficudt to anenge: We have fo distaibure “eu Pastner. because, alt ine out Knovsled is ‘ne 7% oaAmanon aboot a domain nak Can be Used to Solve problems in fat ASR? LOE I= ONS mnony Problems prequire S anv cl Knowle age ond Ris krmowle dge mush ve spepaesen ted tn tne Compuker- aa kmovoles ge repseseniaion is an anea ynrelli gence anak is Comceaned wikn now knowledge about tne Word Cary be nepsesented Gad What Kind 6 pugasontn Com be Aone With that Know'led Sas nepsesenia Vion ghovdd be + - Flexible © tt should be Compact £ nekwncd + fs close to tne paoblero as possible rea Shovid be Cas A Swoik change in me problem should nes In SrcdQ chom ae im Ne ez? senlaion of ine paoblem.P A 9 HEURISTIC Senrcy + 5 Hevsistics ha 3 Bei y ieeace Psy Oo ae wdle in Se anch Strat e7ts 2 as 1S veel im Solving +t \o. Uy Probl ero Wric) ~ Could not be Solved bY Any otha Wa . Solvkons Jee vary | ‘ long ume to Cows ¥ fs Some fran le * Ce are 4s vane * a Wevrishe Seaacn technique He Climbin + Becsk-Fiast Sebkch pee re PRrogRAMMINnG LanauAaesS + a, es paogaaromi™ lang vages has abifity 4o thanipula ar d screws rs ee exploit Knowledge 1 Concepts bhtouq aa Hepresentation Sy enbowe Stay CHAS: beau ae provi ide Expr Omen eet: > AD ee ee pecan ie ‘OA of peal wate unacaled ge Some poplar AL Lemguag °S Lise PROLOG He At HAROWARE Ax vaad wore Mca inclune ange followings: pe ed sees + Lage Een of Ram a Lange Stohage Copacity + & 4008 graphics caad wot nigh resolv hon econ ond ynanipulahor ONC a poe —> is wages | ound Cards 3 mmicnophoneBORING Teste oo eaeal ‘450, ALen plgsian Wrote an Aalicle Cones aaa in Mind rnaqgazin e gives = Mew Alwechon to,déne C, hey. pieces UEP sage. e ———? Trinis arteLe he Proposed a 4 Imitahon game tohich was lar Known as He Proposed tne - raetod for dink ? ame named Turing test. deleamining Whetnex, GQ machine Can —— The Turing test measures tne Performance an alle¥edl inte Btigent ma chine against dnat of a human being. —s» “Tne test places dhe machine anda homan Counlerpart In rooms apart fromm a Second human bein 1» nef ersod to ag te Intenogater . aac) Dee is not able do See ox Speak shay to inex of Hem, does not Knoro Which enkty™ ts actually dne machine and ma Communicate Witt tem Solely by vse =f textvaQd device Such as aA teminal- —— > The inienoqalor is asted to dishnaquish ne Compurer frora 4ne human being, Solely on tne basis dreia Answes do questoms esbed over. dis device. Tf te in Fesrogabor Cannot Aistngquish ane machine (om me homan , men, Tearing QA4gvueg the Machine mas be s8.umed So dy : pe be in Feldigent At eek [Es | I Jes Bem 4 eo 3 | Int SRROGATORReviseED URINIG Resi ee fees Se merkoissa Od feck 1s ce (ua test in uanich tne Holes of A hounran have been meversed evised Tusing test fellows tne Same kegt> © Human Subjects re ne Convasstone style arn: deriberately ig mornings oo Convusetiond Corapurns an oh oano$. SX Amar AS & TOA creme tS to imike a Conversation prog —> Peng. this Loetd mvolves +o Sone degrees ne meaning of Hn that ig Immediately Apparent #o % humana and tne Simvsatton of Me Fonds o(ace oe eae Conversahone? progsams typical make. — 5 An informe sevuse ae Involves a7 Ghrempt fe Qivnviale @ Compu wines tne formed Strychse of tne quting fest. Tne judges o{ tne desk ane Fyplodty Nok AXsnre yn advance mak 4 NewUse Turing test -is ocevacing end dne fest Subject etre phs to edrtcik foro Mne ‘judges! a Kespons é along ne Lines et ts: this areally a human?” Vereen mR eae Ua de al el ie ae aUNIT -2Expert SYSTEI = An EepenaSusice System i is a Cornpvter Prog} : to ack as an “expert in a panticolax domain. Exp. ort S skews ane Knowledge Proqiams Hat Solve Patan em asters Hequisce Considerable ArnounkE of technicak Expestise. The definition of Computes, Preyom is Algositnm + Baka Shuctuae “Proquarn nom Aesigned at ie Tn an Expest System, the de. on is > Ease Engine + knowledge =_ Expat System | Fa " Bn Export System is a Com Prd Progsam that Simulate pene § behavior ae ca howman ie on oxrgoni aioe that has” Cxpaxt Knowledge $ Expuieme ina eae field." > J Components 0 Or Expert Sysxem +. > E ae Sen >» > > USER INFERENCE Maes 6_DGE ITER FACE Evxone Se PKnowLeDse Base > Ps The Knovoled edge Base is tne ete ee €x pont Sygiem. It Contains tne déteiled knowledge om E&xport - Ib ts P Geral im he antes aed 1 Sedge ae also Contaiing {acts , p Qveshin Ss olnen anes a Ps The Core modvle Expat Susie ts 145. Knowle dae Base. Tt is a voane hovse ha Knowledge “Captured {ior Ho human Txpart+ 1 oe P_. tne Capatilines of Expert System axe depenien! on AN > omnovnt of Knowledge tr Contains. We neque lo of Knowledge , to Wild an Effective Knowles ge Base. j Several metnods af Know edge nepresentation ane as follows + Rule Based Systems + Frame - Based Systems aeFENTERENCE EWSINe 07 ——» Ina wvle -based S$ ‘ % stern ne infer: ‘ fi Rude Iniearpnelnr". Ds {enckon is to nee Satie Vee Mikes GOA OM tle vse te svles an, ikemsg in tne Knowledge base to Sol i ene Problem Sek by tne Wser- ¢ eas > Inference Engine ig ne ' REASONING’ Cornponen* Expat System. “Tne word inferen ede ig moxe Accvaake. > Iofeence Engine has no Und tanding of what tk fs eing SNe: Anference Emaime decrdes which “Search Lechniqves Gre yeed to delamine thom tne goles in Ane’ Knowled hase are to be applied to ne problem. de e@ennnaneonoe ens ——s-tre sole of inference enaine is do WOAk witn Me available inf ation Contaln 3 in me Knowled: base to darive new \nfoamanors abook ne Problem: Useg INTeREACE 7 > An Expat System needs &O Commonicaks wilh ne ouside. Wald: ‘ : SS gone Vets interface vo Qh alors ; oases que ston of bene user # accept Aeplies. sete, Sys Wrong ASW. » Me vse to Cosmrect AO Peepusuados ol ce hele: + Hne System to \nfoarn, Bne User of meqatve Cantloetons pre VE interface provides Vaatovs * 5 « 2 e € ¢ « ‘ a * 6 ¢ Belov oo days > - @ User Such “OS menvs, aAapnic 2 ‘ ¢ * ‘ fociWes bo ‘ : ineexface Oye AD ynake ne dialogue user- friendly, Ale ise he snesponsi bility of Hne User. ‘interface horn KS internal nepsesentaHion to Convat kne Aviles. ko an VSerm undaskatable foxwn.Features | CuaracteR istics Or Exeert System 3 High Perfoamance + They should perform at Ine \evel ao Noman expent- + Adequate wnesponse bime % “They Shoutd have ine abUike Hme- Tree fs bo mespond in Oo Beasonable aynount of Crydal especially fox neal time System s- 3° Rellabiiey + They must be weliable $ Should not Crosh. == a\seypuld be able-fo sCaplaic «the Sters 4s Undergkandable + me neasonin process: ty Should geskify is Conclvsions in me game Way % human —Eapeat Explains: 5° Domain Speci tciky o£ Expert Sygtenns ane Fai pirant vo domain Specific. 4 6 Use Symbdlic hepaesentalionS = Exeuk Sys tew use Syrabolic nepresentation S for knowledge (nules, pO Sess vod tamn Com oN iam Tac tht cea ae in Perence thiovy h “Symbolic Com purat ons. vages | Expert L Specie 2 Raognamming Lang 2 SysternS one typical Walken In Speciok pectonatn lanqueges: Tne Use of languag es like Lise and PRoLom in Me AeveRopment ef an expert System Stnplifies tne costa paor ess: e Justi fied Reasoning = atnig allorwes te User oes oy Expat System ko jusity tne Sot vHon Kovtsed ths y tte ae oak System heute le Ved easy do use a: Ee ertat event a= MONE Cornpurer VS. ind 1h Easy to Uses ° An\ Capability = © ur Systems owe Capal \e 4 yo Explaining, Copabiliby Cxplaning Oe rani Geis was neached'EXPERT SY Simee alee ey Gale UE a oS Ge) = EnGi wet z ra Hl Slege PRoBLeEm ¥ study charachaisnes tne “problem Be 2: Study Scope Extent o eames § Hesounces need \pentirication 3: Amount a DECISION ABovT THE \r whetha, Language of MoDE OF | tool ppve woe e 2+ Characaistes of tak Stage? } Concepls needed {oh 8 2 DEVELOPMENT kage 3 oF SelvHon & PRotoTyPe oe tered of Kno wledge : 1 Adina inka.acHon bus smolliple expens 2 Heawy planning dates DEVELOPMENT OF FULL-SCALE System Stage 4 1b Basic Hesoure Kequixemesnt 2+ Pasallek Convesion § tes Hn: 3: Maintenante of Knovolect ASS tn Security of vovious sub system 4 saenbificavion of Ine Problem % An ais Stage, He export f me knovsledge engineer Inmuact Jo ldenk{y +e rxoblem. Tne major Pink discussed Hes of Me problem axe gtvaied - JmeLEMENTATION MAINTENANCE 8 % 5 ARID E voLuTion e for me Conac Heals ees scope f extent Ame problem: Tre arnvount Resour es ioe ded aprcrsomml eg iNe problem which can sooen *rouble ore identified - ay e. ee Decision abovt tne mode of clevadop nent + Y— Once tne problem ie idennpied, meitmmediate shee Urouta be. to Secs \e cavoun “ene. ode oh er need eae etc knowledge Engineex Can develo o 4ne Suche NR Protrammin g language like PROLOG Or, LIS POR. on Comventonad lawhua i To amis stage, Various tools ane ident fied 7 ane lyzed {x the Sulability 3+ Development of a prototype + ——» Whal Concepts ate needed if oduce tne Sdlvtion. one impodtent factor do be decided hare is Me level Ap Knowledge. > Aft is, me task of Knowledge acquisition begins. the knowledge engineor. Me “domain expert interact {reg vently and He domain Specific knowledge is erbtacled: > once Ime Knowledge is acquised, me Knowledge eng meer decides on ine medhod of Aepresentation- —> Wren tre knovledge wepresentalon Scheme and ne Knovled is available, a protoly}oe is Constweted. Se Developrnent of {ull-Scake Sysiom — ee do te full ScaLe implementation, Sub—System vse ot Gantt Cholt , PERT ax cer technigues are Wiekkome. Ps~ Implementation, Maintenance § Evolulon + This is the fina Wife cycle Slage of an expert Systeen- > The basic nesounce AequitementS cure ‘foie filled ecm Poralle® Convesion ¥ eed techniques cue adopted - “me fined System Undergo iS Eseling larer. hand over Eo kne usu. CatecoriesS OF ExeeRT Syvatem + WOE UYCEUG a 4> DIAGNOSIS = Tnfeas Systern malforichions from observalion. —> Medical diagqnosig was ONE tes Piast Knovoled: anea to Which Cxpet § ysiew technology toa Applied. 7 > »~ P 2 Zz 3 >