Gteocluctian Aue Srl bjence + AT 23 > GP fs Payrems Hacktre 2 Hk, analy ge,and make asdecForn Yoselfre bchoucs ke, herman beiege Best Brasrople of AP 2 fo make the Computer Sygr0, 8 Viokiog GelF clsP¥ng Cox8s AG Can be Cass? feed tte toe types ’) Based on Copabiley 2) BaBed or ‘unctional ty A D Based on upabslty LTE gap caleyteed Sob’ S types a) Newtous AP/sigook AT? Free} o> amos AT 28 allBo Callled os Les Ar : »5 phases AP Gonot Perfor bayond 168 Feld Bl lin Fations OL 78 trolned enly jb Spek take “Feomple Phygieg: ches® E-Commerce : » General) AP! be otyped ‘nbellence tohech an Perfoxn any task ith eféency Ve © foment : > OF boskes & fot d ef fels te develep 2) Se h®! oh be fped Fale Iigence eh ean Peto ony teak exrth effidlency more than henanS : Rawwo>) dedteoked for as one tor k Genera) ) Tectoern Pre AT humor Tore snke\yene than hemor2) Based an fanctinalty Stes egetn Cateseteed ko Hype @) Ren cbve Tockhes + Reactic maehdes aye most badsie ype a Arteta Satellence- > ReaelBe Tlach?neS do nist Steve memories n Rast experance Fer fotore Seope, Tb Stores aq Cosrent 8?fuat n. Dr Goggle Abpho- 8) LBathed Memory 1 Lirthel Memery machhes Stove Post Papedience a doto- > Gra machtne8 Gan Store datos ob Iinited tine Pettos: Gaoomple + Self aleiving Ge - 2) Teach Herd? Teary hel can UnderStard the horran @motieng , henon fee Rag } : > Ths whe 4 Hachrnes ae ast developed ab. o) Self Ausene ss + os > Self Assareness Othe Sotoce Ag fed [eellgence -) TheSe cracdkheS well haue moe know ledge Qugaved te homonS-° ob tao Sobset, Arb RED Intelltpree hod, uS25 Sable!) Leowrhng allpithms to butld Systems Hal hwe the nbilityts auto mattally leaxn and Fmpaove the erpeiiences wPthout being explierty pregame > Hb uses Listofeal clate Jor tore DSC an antioduced by Arthors (*) Somuel 20 195% wikia dy ME r olP ° Fades J ML, | sl Otes ‘Pp aoa Algeithe oki? | ee ‘Deep Leartng ; 25.0 Sublield 4 nue tot deal with abies. 9 Deep Lear GS Deep Leacrfing pe ude clep rewall nefsosks to tra a model and ?t automatically buthls.a hiewrchy dobta xeptesentalion. Need 4 Deep Leowntng @ \. Trowess 4 hege armoont 4 cate 2 RerFore Complex allgoithms. 3. Best fer fownance W?th hosp amount cata. {. Feature extzaction. Application s 4. Cancer cletection Q. Auto nomoag diag C8 4: Hoste be Frans latinsd ) jpt® — o> Sopa outpub

Fs slp Sthin Poconns, adel on the Cuttert Slate Ts Envtrenrnent > Gycert Stote Fthe goes also Called! os Eautsenment: 3h ‘Ketnforce ment leo reg pe can clirectly Cerne wipe Mag! > when the acken’s pocdermed there Ll} be Cho aU En dian ment: . ‘cet Ortie Chess. Re a egitenert vtec ae) Fenabty Chis Geputore. a “RewardBo sade 4, Nachine Leav Leawcng > ata cellecton joe ena: have te Gllect the a fs0r0 ab afferent SpoveeS ike Fntetnehd Foutiealar Brgarvent on here ray 2 rey not acgutve colo > Grgof Retent chta Ae a fof tive 4 95 trained U5 iSth clata etd Joo moos better potter than aby bab B wads tered With dataset 4 Jooe le teat A So B becomes , Sul Ree als aa ows, Se there 9% Sn better Peco cenanC? >) Voor jpa lt qfdata values ard Tclevact OF @ cbta Set acguved condkating 4 1809 4 chla- chia called os “port i?2 welevant dota D wantlo Paediet the weafer sreportd, netboloy Se F grease date set % Past loo Yeas foom Fokemet buk get the chala Gel oh Block mrasket iokth % nd rekvenk to the date > ovefibtng Fle arlgttlun Cosel 4 bobies ond bah vofance #8 knsonot Over ebting > Under Ritiog Re afaebtton Gnsbls under Poting > Softioare Snbeqnation To irk challentag ting Tnclecl ping o Sefteame % cabled BePlioare Bhezetfon. Tabgetn meant ombpy-the medales module eLJIThe gual 4, lear oteg, ore understand ard pre dfetion- > Tp eandag Prcler Conia fife Bey the Foret onthat maps botscen the ypu ond the Subpat, Suchet the Leow Pag Foren Con be v8ed 49 Predict Oubpat from fulore Input 8 +) Sbuctvb Reuallee?, 5 >) Struct Bog ee aleceg clk are Inaportont oapect of cla SeFence- > ulhen the goa) ®t Sokewpeel the mode | and qpontily otheuncertainty 4 2m the Aakor, Ths analyst Py usually de Teaxedl to 38 SboRS Real) learning: ‘There axe nay ocly trong Jor oolong datas. Ge lo aceuvalely Rarecleet Some Furbure Quanlity Fh 4 chlo Pokereak hen Boene ObSeaved lala. 2) Se cPacoer w8uall() tobeveshog Faltean in The cbalia- >*b Mp achtee wae eal, One most xely pn knowhelge Bom 3 Jonpor baat wl las p the maithemal? col dcence Y TFonefonal) apo? matPon 3) Optintratt i 3) Parobask?l Phy and slotist cS Fonelton Approtimatfon’ N onathe malic al) Fonelon 38 uSedte Xeporesert the Stelatfon Skip between the UnPables aS a clola Sentish, Soe need! te ordlerStarel how bast te @ppo% Crate and Ye sepatesent Sonclon usteg Jeast amovot 4 Graputer § Porocessing Fmemory. 2) Opting ation * > Gren a ot, revfpernafical) meclels 22 wish to Fd bast 5 wtde! % that deb. - Oks Step weed gies how legge ed optim’ Calfon “al Pthen ancl of Re Pnb Gonpuley cong DFomvamnsig- +S)foubolslly £ aatates The knowledge of Porobubily 4 Slatielkes % needled ¥4 (a) tram the alot and te apremte amedel. fo) Saperiged ond UN cvided Learning wei ; cr B to poredtch ~) Qrenan cing 6 ere 4 bie main gals diroe: Bre Poediel ao patpub vatioble gf! ; : ere les yma degterd Spyabore St whether the Syratore ig geosre @® Qs. 8) ni Fare and clang the aike the Asgcsttoral : : Jand.- - 3? The bags Filed Sthe Bee > Nhe dato Seeabet aabterspt at ts prediction ey encecled {2 @ mathematical Fonction 4” o,Wed the Paredection Sonctian , tokich take an “opub & and Bubpub o guss 3X) bey’ denoted by $ > Sh Degesstonvpareblem, the MeSperge Vavalale y Con fake ony peal 7 A» D> DWCotvack rohen'y’Can boly ReFaa Sintle Set Say yed o,..--1,-e}-then Prdiclig & J 53 Conceptually the Sarre 03 clashing & The input Toto ned) c! Calegoies and So QPaeckelPon lbecomeg a C bas hee Lipe S Problem 7 yrD Le meadue the aceuragy fe Paredielion F Wilh stespect fo o afven ke&pon Se y by using ome lose fonction loss (47) Tt % oolthely thot ay masthernatical finebion’” 12ill be able fo rmmke Porediclions for all possible poor 8 (erg) One ray enGonke In notore “sve menor etfs %s Hat» even wth the Baroe Frpu *) the Gulpat “yl remy be al ?fhesert, cleprenal fag on Chance Circomatenced 4, srancbm ne8S 2 fortis steaden Ne adlopl a probablat aap reach anol asgeme that each Pade (v5) *8 the Out Come da sarclen Fade (ts) thak has Some Jeinb pooborbilityctenstty fa). > ene asses The predictue Perfomance via The expected logs, USually ‘Called the atsh, for 9: for = & Less (9.3) Tratning aval ee ting logs Crating Pater The nog fr8lances A stoi used t aS the algesttim 25 Called a8 Teotnng Dato. Cralitag Loss ! She 1033 Occur! dusting the evalluabton d a creel / Algattien by uty Crating Tate. “Votaset a ot %3 @ Collector 4, alata generally ArUed! Sate Fiakning Data ancl Tes #3 Dates-Festog Date - The no- A InSbances 8 B08 Urged to test Al evaluate the machine lear weg ally SHom 9s Called Ferg Sealtn es ohta Testingess The fas crewed Gog the ebation 4 wreck by using esting Data % aed Tako Jog & > Shemethed ude by Sore © Bell amoont Tpamont infer outpoly Tb algo then te poattl a ws the cup 25 590 §0 Giainiag Loo 56 650 bo Joo 6s X50 qo lest a a O:)¥Soo42 3 << $50 bo : Joo BS : $ 50 qo gtd you Glen Data 2 Or} 580th capamwont [Predkledout |, 54 g 3 Sd row Of k 660 +2 be ath you Ofxgsot 7 sthyow Oly Foot? qe 6H now Oly AsteCTesting chss cBth ane ox bo +2 = 8 ath mo ol KBIOTe = 89 gt moc ok 102 +2 24L oth 00 ‘mle Wote = 5 Total Trasneg Logs i r ep drd, (oss ee ‘S sa ssaree att? 2 bus b.. . C fobal resting Lose 9 i 018444242 ~ BUS »f, See 4 n Total Loss = ob etreeintng toss +o2 Seating Less 645 thd = 19" rradedfs tn Stat tstieal Leavrdra | onc Etta “A rade ipso Stak Lang | gdove Fiteg Peo ie ote cence between the Predicted value and actual value % ColPed 08 bros eink & eats less Valance he Hh afegthrn woh le pedowoing evallrct Bn Shows cod acouracy Pith data Set TE was taSned bob nol Same worth DtherdateSet th?s Fhvaklon we Call 9 as VeonCe We Exhe Pecoenane @) @ dythe 21. 2then varies ba Sed andalacel: Examples ” fx-2 : ‘Pallamvont Mp armant Pall amount | Tp arrant Soe aes $50 so 600 SS 650 bo Foe 6s 450 Fo Oo As 50 te Goo es aso Ae TIL A depStthrn nob Good MN, ub U * LAG Rn Bee 7 Jes foeroance. onder fn A enachine Leaxng Mace which has hob Bras and fms Lgvarce % called vndey Fttieg Cueckbbeg? A rth allie which hes low besand Bh vaseance Qs callbeel ue c Fling. Toe off! The machine Leawting Lp Sth 25th 00 Overhttag Zonder ag % called Trace oF ’ \ Dnderktme esplerodels Jrcthoda the alten eleels the method a.rmang the molfple re bhod forSen put Gachad that the Wisk 2 minirorn Tes Gocept 2% called os Fmpoareal Rak Heaton Oo ° Meey Llaesng dlesathms don be wethen % os facen ds oo opbione ation Paoblem eth abjective te ries Des JosSfureen FO and a siegolafeer ©An allgatthen hag toe methods £60) 2 F (is) ro he Some Taput 3 passe) Paoided to bot, metods and Oubpub paedielel srespeckrely 28 Shon bebo Nolet My Paidoh % 4 dhovkd nok get Hed From the asbove two Poed%ebors 4, methods Fe) @ fol) pe Select method f(x) as poedicted the outpus wth pereno, CmOK: Yachine Learn’ag eo make Computer Lean be need Arefed) Neu wall Networks val) pletionks Homan Leacting cabled as mala) Learning = New praked homang tp leasn Aseftcel round oeloes's * Gh % celtng bab a Gllectian 4 neviens Gnnacled by osing Some nebirosk8 Every nevson % able #0 Commudicate path ethes neosen % the nebwosksSapeb Hidden i od Ske ait heal) al nebeoorkS generally onsets 8 fayers ) Sinpet layer © The neuvons present Gn the Inpub yo yell take the ‘input from the steal) Llow ld. 2) Hilden Layer 1 The neno8 Present 80 the Halebon hoe will Porece3s en “pul ty produce Some valued. 2) Bulga Payer + Te ceouna reset fn Bbpk Boer ull Show the & uk forthe steal world. Fretotall Foput Pporedent sreceued ak nevsen D 2 U9, %, t My g + OF gt brag Che tall Tapa Beceted af nevsen 3 E Ys (ph, + Ws + ars + bios Che total ‘mput geeed of nevsen F fs (0g %, + %8 49% +bia8 4 “The to tal tnpat areccned at nevsen in the. bdden Dove 95 Cllled Leighted Arcrase-Activation fonction Oh 3 nothieg batas mathemadia aD Rirmvla Present the neosens 79 te Hellen Laser ho Nevsen pmecenb 79 fhe Helden a Kecete ji jopel Fron Seput layer X — > ° ley 0% FW Xy PUI +08 x5 +63 6 Some constant, Gy above Exemple te tall oped Se cene Sroen the NeVsOOS (9), H9%9+ 03 84 b : Z Wor + b; Fe rlteratieal Joule bake Yop valle and Paodece Some values- oh mesh ll v eae oF} Bo © Gialank wells Gropoved) wth v ve Poredlacecl by actialien fonclfen os Batpat: Bs ails Ppandocedl by Ackvakon forben 5. 13 ess thew, Cileciatt value the cotpad & Eerold)) Uice vey &a-Applications of Al & ML and AIML: 1. Gaming: In any game if you are playing with a computer that indicates the computer was thinking like a human .tt Is possible only when a game thinking only Al 2. NLP(Natural language process): it is a module present in Al whose specific task is to translate the sentence from one language to another with same meaning. 3, Expert system: Expert is nothing but whose gains more knowledge based on experience(or) intelligence. A machine installed with Ais treated as a expert system where we can cloar all the doubts we have. 4. speech recognization: convertion of speech (or) voice to text is done by Al. 5. Robotics: Robots are two types a) reprogrammed robots: this robot was created only to do a specific task. Ex: cleaning the house b) Humanoid Robots :These were design to mimic /behave/pretend like humans. 1 It is the early humanoid Robo invented mainly for the control the human emotions. conversation between the human and Eliza Eliza: Hi User: Hi Eliza: what made you come here. user: My Boyfriend made me come here Eliza: what's the reason? User: He felt that |am depressed, Eliza: Are you really depressed? User: Think so Eliza: In what way? User: Men are all same Eliza: In which way? user: My boyfriend pressures/forces me to do work Eliza: stop talking because Eliza stop taking because appointment time was completed.G YladSne Learomg _—— ons qow= ins Q wv Day | Dib look [repent [hl whe | | Sony, _ | ot Ith | weak 2 denny vot nah Slvr | 3 | pvercost | bee Hegh weak | edn |. mild Kah | weak ¢ ‘Rein ean ool hosp} ea Je] 6: Rain Coo} Hore) stor} re brercos t ft: Coo | [resenal sto] g \deany | neld | regs Baa ] 1g 4 | denny Coo normal Make a “Rath wild | Gorman yn | denny cord m7 Evera@st Aw bukes= Fealued: Yaga lesoe Ids Paden Fe olate. oSing that pare ak ple Con Play cfcket on I5thalay Sre Condition are qertperatore < oly poweedety _ bah, 18nd = S900 ; Bat look = Sonny Caplan fe, \4 Sows t stall) : ae Posy 90”) r 5 qe Caplanoton ene But: look . putsloole ea)! Step £ * Ves: (9) Sony |) Seat il Orn (5).. oS hts tel) yest) foe @) yes (2) (3) ? (donny / res) -) Pro alsltyey Sonny co Owe : 7 SQL QI 1 yal ? (sony J v0) Z 3/6 :ee [al9a9 pear (3 ) St10ng (4) yes (s) po(2) yes(3) NP (3) (wear) yes). * oh a p(weak/no) 2 215 o (eters '8) > 2P 7 (tong d= 18 Xx = is ‘ Cool y, Hah (Brorg ve ciev Catectonds “ : aXe ee (seensfe) Xx (coal yes) x P (wiph/ves ) XP (trong /yes)3.3 : 12 X = ta 2M og 24s pa een | ~ 9-005 24 6 (AY) s P (no) ¥ P(seting re eo 7(cool/n) ¥ oC Hayes) x Pst / 8) 3 "Ss we 1 BK 2X <7 ae o> et eee rw Aba hg Nefe “Bayes Al Cone ttonal probabil r