Scribd
Upload
57 views45 pages

Artificial Intelligence

Uploaded by

Reba
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
57 views45 pages

Artificial Intelligence

Uploaded by

Reba
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 45
Se fyb fical Soleo pamhtok 2 decides PRoloG, LISP, crip 4 ae Tsk Domain Al a a eee C@ Mundane Task > Peveeption = Speech © Nip — underctandin Guenevatien “Translation. 2 Conmen Sense eeetle 2 Robet — Control @ formal “Task © Games — chew ~ checkers = 2 RN Baok s. Russd - Retry Novag Act xationall y Sean Rule bacef P-L fow AT G) Expert Task Enginccving Design — fautt india | a Ce Cae # Sclentie analy 2 médical aGiagnosis @ Financial Aneigsis M Taying Test Back qammons - Ge o Mathematics = Geornctey, ~ logic. i treqral calcula Theorem proving Be ewaledee pole asim pe a!) 34 oe ° 2 ° 4, 2 4 s Ei ZG >} iB \ a @ . 5 nena 7 Sheps 1 6 iB 2 & z 2? ° beh” s Sytem tuate toink Nike” Human srk 9 CRT ’ y aks acho » y ack vahonally Sn AL sa fey to build vetional agent ( autonomous) Z 4 Hi eat AL) =s e-bonc ll adineees He Week AL vs strong AL Saka astfical Nard ee pee CAND) : . queral purpese — ( alithien tt te build) s only fav specific use geal) a One > nok er AI * Rosen belt Me Cavity > AL taven develope . * a, 2 3 4 5 20. ae we uo ee — Ga fining 4G dav Co a-d r BG jar 4c Cx y) > Cx-ds y) powing. cut Cty) 2) cnwly 30 ye iy tn 2 Age A pour heii m C4a ce ° gnitial Gnd rah sia Rule 2 2 wea) Rate 9 F A Rule 2 gale ° a gule 12, Bulle 5 ss ° Rule a Rule 15 . HN ae Oe a ey oe te api - Rac) oe raniagee Ee es, an blew ub x Reus ie G ae a greek ja this a prbien Doade +[2[23 pecan Stavehy necqnealeeee ae are Sending through 2 — Problem Salvin * hetion. What Choices Combinations fa nguage Ways map words de parte of Spach / zenlenkce Computer Vision Natural ways te map features to sbject model Machine learnin Possible Concept that fix examples seen co far — Melon planning Sequence of moves to reach goat/ coatinston ag : ChEASEccscrip te = Sencov —— cae ee ae i 1 EI eee \ “ eka nt L fo od . eo + Bane Sieg Mane Problem Selving- Agent ( Percept) eh eee P stale = update stele Cotate percept) Blane Jaynanic 2 if state is empty (or saquence is empty . ce am = formulate Goal ( slate) Problem = formulate Pra lhe ( state 2 gen!) Sequencé. “/Mapecay cine Problem) action = — fivst ( sequence) sequence <= Rest ( Sequence) . return acton tad a State | Ewvirounnent > Satie or dynamite 2 2 Fully oF posey a observa 2 o ovisrete ov Continuous © ° OBeterministie oY non Deterministic C shehacke) @ ° o Episedie oy Sequenb'al 2 © Sgt agent er Multiple agent i AE Serch 22a _Snfermed “Bling Seach “Seamed ¢ oe 0) pts a BES AP aes 0s ee i nn nr 4 ae WD Blind / Uninformed Perfor wmance parameters for algor : C2) 9nfovmed / Heuriehic O—completnes 2 @- ofkmal (G) Blind © /Untebinm drones ea ees @ Tiere cmplexity 2 —DFS uniform Cost Seardh @ spre Cates — Depth Unnited ~ Stevative Deepening — Bidiveckonal * a + ® TTiene Complexity — b i | oF 48 i weap Be ee se Ss bet = brawcking Jocks - For BES eee ce-| Niemer Sept Nodes Time te ¢ 2 neo ts 18 HL 100 ais lo GMB 6 ao* 49m a0 GB @ 40? 34h 116 40 40" 124 days 40s 18 42 ao? aes “ BB ye 40 pehabytes 44 40 $52 3 gies 1 Exabytes F4 ore a Sherative deepening Search 71 > a ce eetaa Hae ; eit” oe Gi: mit = 0 © limit = 4 >@) A gD = A 2® oy ® S limit = “oon So 32) 26) Ss ie Q ‘ Soa No o 30 se uninfernaed Georch 5 not gerd ce as 4, 4 an rite 5 form Get DES Depth Drerative poiaseectre| 7 i na eae | | himtled) deepening Connp ete yes Yes Ne aie Nes Ye Optimal yes Yes No ne Yes Yes ve Time o¢6*)) ee ee cen) oc) (+4) oC D 5 A wy, Pee aes age uM acon] ace) | 02%) | oct s e j ce ce, ak a m > maximum depth ‘ LoD depth Limit e dd > alepth e e e ® ° Ae Sn formed Search / Heuvishle Sench On orn we eas Ga foe Cn) (es tena rade Te eee 745 qc! mode path frm node nm ote > fad optinuin / chepeatt ns =I AE Bet Fivet Seareh as oe wn 2, : oh ys a a fx aos gC OS On, Ore teve | [depth z 41 a B42 qu etait) B43 pe C - 1) ecnj= hinyt Shee: Cathe wach aed) a | eve stienetion Mag be a2 there, These 2 problems may be tre in BFS “+ Algo is empiete > averestnale mot 9 wpunderesfinnte i Gonatat dey se sine iely wg wevely oveveatimalee he by move, shane ‘Seo te Ate 2) a lp mst is more then & canal Sipe veestionsle oath) OPEN 3 Nodes have generated and hove had the heuvistte ge applied to trem but mot PKamined yet Eiosep z node have already examined = PME es cae ae 1. OPEN <— 4nifial State 2, Qepat “unt eel Ml eae m orpeN do Anare (5 no noda lePt @) Pree Meibe “ek Onell AGB Y Gemraies, ite succemer been generated before evaluate ity y e Ril Chimbing Seaveh RCM ell QO coe Conecimamon ED g_ Owikal Stale a) CS 6 a] = Ey E S |e = eB Le a} ie Tal he [eR A Tal fel B 6 © @ VA Tal (al ae a mae een ee. 8.6.0.8 © @ a ro Aelia Simple hill ee @ Evaluate tee wittal Slate. Ht a geod aioe vebwrn i md quill) sere: TUwhivatns vel Coca eR IAT cy the cave Slate 2 ) Loep umtil’ a holukin it Sound oy whl there ave no new epevatore — lePte to be applied «Iw the Current State @ Seed an oferalor fat has net ack been appiicd fo the euvvunt = State and anole ie to produce a new state ® Evaluate the new tate Borate ete gen skate , vetara if and quik. Cy) Sie ie Se notes ome re tate , bub better than cuveent Stake then make it curv State ay Sw mot better than the cuvvurt = State, ther eoutinue usith the loop Brae ggse $ Pe Ascent Hit 4) Evaluate dre inital state. Sh itis also = get Pity, the wd , nm velurn it and quit . Otherwise make Has inva ctale % ourtent 6 tale. 2) Loop unkill a 40)" is bund ov unbil o cownplete itera Her voc ee: preduccs no change to the Uurvent state a) Let , Suie ibe. a dtate, such, thataganT posible Suscogevis sph lome leurrurt state, osfll, ke Bern Sica suce« ©) fav each operator het applies to the wurcent_ ¢ faba iia Q Apply the operator and genvate a nw etale GU cyailste tne nel xteten a te co state then qerum te aul gt nett campers th be euce yt he Raided ahen eckeaseeh anc tenia Stl dy this vat better tvcvneue a eagle c) ty me rue vot ell nantes Pelercerte steers Tha co carve ctake- fo suce » 3 eSeenapsite OORT HE: Rik. eGo D-£) initial atelease © EBs) a2 On Gay Boe od: o> Oe > End string Rat Zz > C8,2e-m) aan Mm ce Rt Les Nr me C8, bb) Rs \82 CC DL BLD Cerin) Com, 6.2 = a aM \ @rt 7M, , 2) CG Mim, MM, MMs ay (means meni, @7 7H) | Coyesmm 8-6/4) BA mised Pa A i 7 OLA OSE / @ ( 4 wi) Cp LM My Ms Mm, ) oO @ BLM, mm, M. GB, Bm) (ate fe poset) > Jn cram wee ome wale at eath etep onl C mention ely) Jebel Jette Bee a | 2 2 Qs Vat Swe use Backtrack y Lg Ee ye ee eee O. 8. 6.6 0.6 ©: PV PPE ETT EE EUUFVVGILVODODODUOOOUS AQ - pwasle — provem reir ee rh 24ttl= & ye jitia Goal Oe es 8 [| +! 4 +t\lcec|s 7 { @ \ ees 3) iw dsr | clement] pein From its dew eal es ge] 3 hCn) oval Ca De I Ol i ne hh nd dk on fp _) B. 2 = ~~ Ak Genetic Algorithra a ae 4 - | eae | ae fsa Meme ett Tritial popalation Ay Selection fines J™ Gren Cnvanne mare Geneme Pepata tire Mutation: Cyossovey™ we Rew Slater Oe Generate random population of chremacome. until Rie eee Darerestenncs. pepeistoa: Ae felis @ Evotuate tw Gihnes of cach ehremosome ® Sdeck two Chvoraasaines (povent ) fom a populahine meighied by tneiy Silness - © with prebabitity pe evessover dhe parents to form a ned of ferrin. i ion ith probablicty Py) mutate nen Scan at cath posit Ow the chvanaoseme- ; : ©. plosccnaiiche treacle populahon G)~. Retwn the best ot 2 are \ eS + Twihaltze thet tb plenvant edoton Gand beck squkien PF { Generate Ane sel oh neighbourhood felution So tne Unrrint Solution £ Y find 4 best Solution from S$ baseol on tne objechve YT value cavelas A CO 9S! th betbers thon is 4h 5! co u oe | Delete S from candidate See px Cp) > width eh the Beam BAectniay) OPEN = Rinihal state} while open is vet empty do 1. Remove, the best inode fim ropey OPE Nae CAltaa 2 don is goo! slale , backtrace the path te % Fromm yeeavded part and acta the path « Oe Create n% Suceemors ie Evaluate. each Suceesoy, add HW fe OPEN , and reeard ite paver Sr Ne llereNl> Go dake re eecele aden Ged p) a per the neuritic emove —eancnorneds ter ORS aes ae oo Ss 4 eel eA Qu \, a= | € © Oy. a eee Ne SIN TN OT Oe ee ty ree 4. OPEN = SAY pcre ) weeds Chang 2. ofEN = epee : 4 open = LEY 2 5. OPENS SG ee ee ea eee, eee ee ae Se wt ae aT a ee ee 4 a ¢ . De AON Ge terete Consider lowes C Or) aa © ar OPEN = gap Ores) = Ge 20 SOrEN = Sge OPEN = 5 04 OrPEN= YY oPEN=z {ay Bes) 2\0 fim= 0 aa Se Ae Bet be soley ot Case T 7 bet Suboptimal geatnode. G2 appears in Aringe and cosh of fis eel" in OF Ac Subephmel Giz HC 2) = 0 We Know fC 4&2) = ac a) eh (Gz) = g Cau) Gide Gc cyanea: ing eats a n' ahtch Is optimal CaseS 3 aca det 0" path ib h Cn) is admiasible heuvishe. coup est fu) = qc) + ht) ae OS Se suretictl oo Fa) ST a CoseD and case SD fin) & C* < FCO) FCn) will ae max C¥ but Sil trecouthan ib. hence £Ch) & ophmal Put Ihe Start node 5 on Tuthalize -elesect Closed & a Bos Ce comphy oF geatnede Kas heen ence hed Remove From OPEN a node at which fib nai nin» and put ik om eLosed UA ts node. Sh! BM an 0S goat node, att 93 expend “on', Geneka try all Succemors: Ceep) BA for each succevor on’ of 2A | caleatate £ Cn) = qe Gy) = qn + (mW) + hon) eee id DAD IEW wan weighted on SpEN new Close Umass Ste (ae OPEN Anno Prygetdtan) Go te < rveck i Peo pean, ck toe te OPEN jer er ChOse Die its spornbey” a wth z fue path gelding. tet Vaud. Cig Gre) Sha ancees the. tomate FC) 2h th int equived printer adjustment and was found er ce ets EIS a db a geal mode & feund ‘men exit, the eelution obtoined by troeking lack throu yh pointers: failure . 2) @ G SN Mlcenorp Bounded euvistic Search 1 2 Rees may (@ Tes| A)a06 Benen Taz] Bie : ca) if OQ «4 a DD iut mi & @ a2 [Ag] > eomember oho : 2 oe j © ( 552 Ae m= 3 va Se laut a te owe i) q Out = @ on 44g oD @ es eS <«* At an) ® " Y Ae z G) at i si er: , * at yer si ae Probie) Rasa 2° Unie faiture REFSALGO C Problem. nade , falimt) reharn a Selutom silo fieit 4 alternahve ¢_ the 4° Ae eet St Crore) Mee Pee avin to te e qic_CFet) ov Fivchoxdey predicate Caleutws > FOL deals sith objects, Pr 1 Properties, telatiow and 1 funchion - ani : (@ Bc NIRV ona Die ho" el > mpi eeation| ecompienhon oie (fx as) Greve exist) Constant. ab =o SSS eee) An eles teat Ddiequivalencee Vaevione 9 |X [4 ee PC eee eer e ps padieate aed > objects fn Fated (selma: tea cxceaey) jee even ss 2 os —_> Gace Ive éx4 % John u a man s man (Tenn) Crohn, books A eee John tikes books and must aie. lia Caehn, muris) x4! An elephants are grey jn coloy : eS Z exh! John ues aA “ein ofp Food Be of" 4 2 aH elephant Cs) > mieahas C« srw vn nicest (vena x) A Food Cx) Chae Cex Yhedy ves some bady eo\™ _ sol" WY Fy toves Cx ry) Ex Bs Theres ig. covdesne. nd en Re Simple statements 44 fmplemented > || D Connedtives Predicate logic 1) Representation of wevds i Ce & predicates C je properhs oe ula eves abojecks objects o« veletion abject s) 6 are Bh Quantlicr ured to vepeesent tne of the Sentence Cavincetive — NVa'7 2 2 Qunthas = x a meant Sy Compt ome, Stem A i implemented > ASN EL ; si 7° Pa] eval eae re @ F F 7 F 7 F T ra e v Tv a (= U F F F a u T F rl oT £ g T < validate Struth “Table ean be used tr Define tne validity ch @ Sentence fe Chr Diffeat Th the sentence is frue in every 700 4ytea ot Logical Contraints) then the sentence (6 & vais Sentence. ce “ec (C eVH) AIH) SP Check whether the given Sentence is a valid Sentence ov mot . f 4 ZH evn | ove aze | C@v Ha 74) > Pe G te Tv & F ms Gaal é te ¥ a as e T 7 1° aL v v e — F = became the Sonfante. Wa au Va Ul Ser > The qe in ever om fe clifferent type ob genrence i te Rules of Snfernce for propesitional egie * oe Pgy Seo seat oletion gene me Pen be derived a Gerrenks) fiom oc by toferen ce = omens ov Smplicati i a. lication Elimi Ve oR oe aa cca BB eece 7 Pee a Ava Etiminahione A NN Ps S ot A on RR eno x: ae Cl yA ashe tae cr CGitey, hse Oa ee | Kr Ame Nog Nn FY Ov —Sntrecuchon- ee a ot CGV. ccs Wempus wor an corte Fame pared on an agent ane expire, a cave consisting pee coarse conten Cee ways . looking somenhare — iv the cove, ws the Roni cath | tate eats, anyone who enters its weer, Some eee contim betterniene pits that vill trop anyone ao wanders qrto their oom = except for ND Aeon ce fall in) the. Atk Wee kt find the heap ob qed te ange gnc ob tne EV Envivenment p= chs rg Zkach > wi |lanso P \overa ome SP lice | i Agate | e pelle Pp a a 24 eae ca ae 4m awning doy > DAY ee Eainy) Pay Csnrg) ona isa aa 9 Tomy ablorewn Coley a er Clovay A Color ( bean) AU brown—esler dope are ek “ferociaws vy vy x\ grog oe Mr Ratan Govs erapley eet hes bres salaries A +9 (4) > tHe @ a bank acount . Heb @ ant «emp lege ave depositeal in Saving accoumt - HHS OORALASAMSSAAAADAAAAASAAR wwe a ele Fae Seas S Se ~~ Fr ain g BG Wrile a Sentence dea @ Sunny Day Teday is a sunny Day > Day CSaany) y knowledge Base HX Sun is brags and weathey is pleasant Sen ( bright) AA weather C Pleasant) x ‘ HU, 9 am Rupangkan ang kan > Nowe Cn ao as T teve playing chess > Lave Play Cx, y) “ Films are ao Window inte (ery AAC Freporibinal Magic Grrommae ( @NE) dentence > Atomic Sentence | Gmplex Sentence Atomic Sndene > True [False | symbol Syme o> PL BIR. Comptex Senoue. > | Sentence | C Sentence A Sentence) [ Ceenlence VY Sentence) PQ. nt wea \ ( Sentence 27 Sentence) P>&AaASe 6 Gate <> Sentence) y z He Re B c[oearsicie “HEE CAAA maE fen (CoL/pct ob Wevals . #) meet KB C eucwiedge Bate) and queries must have — disjunchens Ganjunenva Normal faye Cone Din Ve we Beata el C40 Dm Gee A (wrt 1 hore ) K literals per clause ey > P Q Day (sunny) re eee pe oes een een Hove Cluuke! “ugbalend dsjunchan oh | trend Comet watch | = akmost ane “& posibve (aa Vele Vv OG Gin ys vo Clause exacig ane pesihve, Kheral called Aefinibe Claune Peritives \itevals = Ae CAlleauneed Negative y y bo dy pvapgition: Seth ots jortntte Gaue (© renal umn aig Wevale) 3g ot Wares ice cenantee tisch fo on Olejeck Tex ¢ Canskant symilel AT oo) promir 8 4 formed Fen pee ees recessed bane ee 4 a 3 parenthesiged tisk oe terms LS a amme ~ roti. | @ iharae Dehn) Nowvied ( Father (Richard ) , Meher Gasha) x DY Sie e Y ¥Y * Uxry) — a4 v aA Gg Foy) S skoleron function |: Fathey prev /R vol fined) Now ( © C Pickard ) ye (% Kren ( ‘ 4 ow vied my i pie Skolmenization UC mat pen a © AW Kings are perton ee [king Ce) teeple (x)] elcing dohn has a crown on hes brad - = 3x3 Commn Cx A onhead Cx ,dehn)} fe Nested QuanbFiecs == EEE ve Stblings “e . Agra brie © Brolars are Siblings rel Aienship~ — varey § [Broth Ory) & sittings Cx 4) Everybody is loved by somebody = x Sy + Love Cx4) o mere is someone Whe is loved by cueryene Geese va te Love C xy) ae wetp = TJ" P Neate pelt Wa P qax1e dane = syne einlene ee Valles (pa tas due in atl mode , PVTP % Sabsfrabilty 7 A sentence. (5 sakishable if it is qrue fey sone me dels ‘ ti tax 2 Semantic ~ Leaic » first ovdey 4 “4 sees w Ce Meaning. a i Producer re yee Aras o TEE —_— p-1.0 > Cia we ee a a ee Coon we : mother means — perione female parent Cae Sol Ym ye yaothey Ce) = m > female (m) A parent (mse) (2) te roof Hi ‘ i wsband is one’s male spouse - eo weh f usband Ch pa) gm mall h) 9 SouteC orn) OR ae Female are disjeint Gnbegerien a ae Male Cw <> | Female Cr) “ Parents and child are inverse ve|” Hy ees pacunt Cog child cere) OA grentpantt a a gave og once parent 30" > apf grendpavet Cain) e> aa tarnt (qi) ~ parent (py P4) ‘ a oe ‘so methey ahild of onek parent oH xe 2 Siblin 1G: 7 Ope ney 3P pak (Pex\ Ai Parent CP, 4) Exare 4 Sam is a0 Dectov- a All dockevs ave peerle 3. Dame is the Dean 4 (Beant ave profamors Al pre Fesioxs consider eer eee cn Somesne - 6 G 7 fiend 7. Peaple ony Crikeize peeple that are mot aneiv friends & Paci : Sam. Cvikzed Dawie Aa 1. e_ doctey (sam) APH 2° ub _ doctor Cw) > Person w me is_ dean C Damie dean a fviend ey dowel nea him: ye We 5 [is dean CH) > is professor Cd) Ss evan cy [Fy Cit pre tenor Ge) n is-deau(y ) S,coecenl Cw 1 Know i) 6. A (By te tviend Cy, 0) og Cy Cie-peron Gr) A paren ty Jn) evtive Greg ) CrtHeired ( gan, Damie) Tic Friend Cx 4) Query = ds Dame no fyiend of Sam 2 oT is tvient CSarm ; Danie ) AE KB in Machine Representation ae PL Ca) WC PiCx) > 03 Cx) 94 C8) 9. -¥™ 1G PacHye yeraicw)) vn Cre Cpa cy A 04 OH) > Peery) v TPs Org) 6x Cay Cm Cyn) F. We (4 030%) A FsCy) A Pere) Ss Ten Cra) 8 pe Cxr4) Qurtion > TP2 C BA) w ah caaag Knowledge Englnccrin 9 a aa eae 4 ena probes (4) Taentty tn Task . Ca) Assemble Hu relent Kuowledge- Predicate ,Ftunchens and constants . query (9) Decide on Vocabulary of, A) Encode General Kuedledge abent the Domain (2 Encode g dureviption about the Specific problem thslamee - () Pose. Qucvies | fo) tine Suterance sytem & qt answer: @ Debug the K-B - 4 Godelg heovun A Recalubion proof ov pre br Regulation Convert to ONE Pe 4 Pe e a Rk Rela eee os al aPpVva QrR TAVR To , pve Pe iM q a ig & ae ra TR ae NY ea \ 4 Vix feather (9) Same res, 2. Feattays Tooeety ) Birds (Tweety ) 3. 16 Ree a Bird 2 ays 4. 71 feathw OY) V BivdeC 2 Feathers Cucety) o. Bivds C Teecty ) @® @ so Assumphion, —| Bivd CTecety ) is False a ee = FeatheyS Cimecty ) = Sey Tweety 1s Ber o =A ca Prong lop yeSoluin «

You might also like