{"id":53222,"date":"2019-11-26T00:00:00","date_gmt":"2019-11-25T21:00:00","guid":{"rendered":"https:\/\/prohoster.info\/blog\/blog_prohoster\/grafik-gartnera-2019-o-chyom-vse-eti-modnye-slova"},"modified":"2020-02-18T14:01:06","modified_gmt":"2020-02-18T11:01:06","slug":"grafik-gartnera-2019-o-chyom-vse-eti-modnye-slova","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/novosti-interneta\/grafik-gartnera-2019-o-chyom-vse-eti-modnye-slova","title":{"rendered":"Grafiku i Gartner 2019: p\u00ebr \u00e7far\u00eb jan\u00eb k\u00ebto fjal\u00eb moderne?","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

Grafiku i Gartner p\u00ebr ata q\u00eb punojn\u00eb n\u00eb teknologji \u00ebsht\u00eb si nj\u00eb ekspozit\u00eb e mod\u00ebs s\u00eb lart\u00eb. Duke e shikuar at\u00eb, mund t\u00eb kuptoni paraprakisht se cilat jan\u00eb fjal\u00ebt m\u00eb tendenc\u00eb k\u00ebt\u00eb sezon dhe \u00e7far\u00eb do t\u00eb d\u00ebgjoni n\u00eb t\u00eb gjitha konferencat e ardhshme.<\/p>\n

Ne kemi dekoduar se \u00e7far\u00eb fshihet pas fjal\u00ebve t\u00eb bukura n\u00eb k\u00ebt\u00eb grafik, n\u00eb m\u00ebnyr\u00eb q\u00eb edhe ju t\u00eb mund t\u2019i flisni n\u00eb k\u00ebt\u00eb gjuh\u00eb.<\/p>\n

\"Grafiku
\n
\n<\/a><\/noindex><\/p>\n

P\u00ebr t\u00eb filluar, le t\u00eb themi disa fjal\u00eb se \u00e7far\u00eb \u00ebsht\u00eb ky grafik. \u00c7do vit n\u00eb gusht, agjencia konsulente Gartner publikon nj\u00eb raport \u2013 Gartner Hype Curve. N\u00eb shqip, kjo do t\u00eb thot\u00eb \u201ekodra e entuziazmit\u201c, ose, thjesht, \u2013 hype. 30 vjet m\u00eb par\u00eb, reper\u00ebt e grupit Public Enemy k\u00ebnduan: \u201eMos e besoni entuziazmin\u201c. T\u00eb besosh apo jo, \u00ebsht\u00eb nj\u00eb \u00e7\u00ebshtje personale, por t\u00eb pakt\u00ebn merita e t\u00eb diturit k\u00ebto fjal\u00eb ky\u00e7e \u00ebsht\u00eb e nevojshme, n\u00ebse punoni n\u00eb fush\u00ebn e teknologjis\u00eb dhe d\u00ebshironi t\u00eb dini trendet globale.<\/p>\n

Ky \u00ebsht\u00eb nj\u00eb grafik i pritjeve publike nga nj\u00eb teknologji t\u00eb caktuar. Sipas Gartner, n\u00eb nj\u00eb rast ideal, teknologjia kalon ndjekur 5 faza: nisja e teknologjis\u00eb, maja e pritjeve t\u00eb ekzagjeruara, zona e zhg\u00ebnjimit, skaji i ndri\u00e7imit, dhe platoa e produktivitetit. Por ndodhin raste kur ajo fundoset n\u00eb \u201ezon\u00ebn e zhg\u00ebnjimit\u201c \u2013 shembujt mund t\u2019i kujtoni leht\u00ebsisht, p\u00ebr shembull, bitcoin\u00ebt: fillimisht arrit\u00ebn maj\u00ebn si \u201epara t\u00eb ardhshme\u201c, por shpejt u zhvendos\u00ebn posht\u00eb kur u b\u00eb e qart\u00eb se kishte t\u00eb meta n\u00eb teknologji, kryesisht ndalimet mbi numrin e transaksioneve dhe sasin\u00eb e madhe t\u00eb energjis\u00eb elektrike q\u00eb k\u00ebrkohet p\u00ebr t\u00eb prodhuar bitcoin\u00eb (di\u00e7ka q\u00eb sjell probleme me ekologjin\u00eb). Natyrisht, nuk duhet harruar se grafiku i Gartner \u00ebsht\u00eb thjesht nj\u00eb parashikim: k\u00ebtu, p\u00ebr shembull, mund t\u00eb lexoni m\u00eb n\u00eb detaj artikullin<\/a><\/noindex>, ku diskutohen parashikimet m\u00eb t\u00eb ndritshme q\u00eb nuk jan\u00eb realizuar.<\/p>\n

Pra, le t\u00eb kalojm\u00eb n\u00eb grafikun e ri t\u00eb Gartner. Teknologjit\u00eb jan\u00eb ndar\u00eb n\u00eb 5 grupe t\u00eb m\u00ebdha tematike:<\/p>\n

    \n
  1. Inteligjenca Artificiale e Avancuar dhe Analitika (Advanced AI and Analytics)\n<\/li>\n
  2. Kompjuter\u00ebt dhe Komunikimet Postklasike (Postclassical Compute and Comms)\n<\/li>\n
  3. Sensing dhe Mobilitet (Sensing and Mobility)\n<\/li>\n
  4. Njeriu i \u201eShtuar\u201c (Augmented Human)\n<\/li>\n
  5. Ekosistemat Dixhitale (Digital Ecosystems)\n<\/li>\n<\/ol>\n

    <\/p>\n

    1. Inteligjenca Artificiale e Avancuar dhe Analitika (Advanced AI and Analytics)<\/h1>\n

    \nGjat\u00eb 10 viteve t\u00eb fundit, jemi d\u00ebshmitar\u00eb t\u00eb nj\u00eb ore t\u00eb shk\u00eblqyer t\u00eb m\u00ebsimit t\u00eb thell\u00eb (Deep Learning). K\u00ebto rrjete jan\u00eb v\u00ebrtet efektive p\u00ebr grupin e tyre t\u00eb detyrave. N\u00eb vitin 2018, Yann LeCun, Geoffrey Hinton dhe Yoshua Bengio mor\u00ebn \u00e7mimin Turing p\u00ebr zbulimet e tyre n\u00eb k\u00ebt\u00eb fush\u00eb \u2013 \u00e7mimi m\u00eb prestigjioz, ekuivalent me \u201eNobel\u201c n\u00eb informatik\u00eb. Pra, tendencat kryesore n\u00eb k\u00ebt\u00eb fush\u00eb, t\u00eb cilat jan\u00eb paraqitur n\u00eb grafik:<\/p>\n

    1.1. Transferimi i M\u00ebsimit (Transfer Learning)<\/h2>\n

    \nNuk e m\u00ebsoni nj\u00eb rrjet nervor nga e para, por merrni nj\u00eb q\u00eb \u00ebsht\u00eb m\u00ebsuar tashm\u00eb dhe i caktoni nj\u00eb q\u00ebllim tjet\u00ebr. Nd sometimes k\u00ebn\u00eb, p\u00ebr k\u00ebt\u00eb duhet t\u00eb ri-m\u00ebsoni nj\u00eb pjes\u00eb t\u00eb rrjetit, por jo t\u00eb gjith\u00eb rrjetin, gj\u00eb q\u00eb \u00ebsht\u00eb shum\u00eb m\u00eb e shpejt\u00eb. P\u00ebr shembull, duke marr\u00eb nj\u00eb rrjet nervor t\u00eb gatsh\u00ebm ResNet50, t\u00eb m\u00ebsuar n\u00eb dataset-in ImageNet1000, do t\u00eb merrni nj\u00eb algorit\u00ebm q\u00eb \u00ebsht\u00eb n\u00eb gjendje t\u00eb klasifikoj\u00eb shum\u00eb objekte t\u00eb ndryshme n\u00eb nj\u00eb nivel shum\u00eb t\u00eb thell\u00eb (1000 klasa sipas ve\u00e7orive t\u00eb nxjerra nga 50 shtresa t\u00eb rrjetit nervor). Por nuk \u00ebsht\u00eb e nevojshme t\u00eb m\u00ebsoni t\u00eb gjith\u00eb k\u00ebt\u00eb rrjet n\u00eb t\u00ebr\u00ebsi, q\u00eb do t\u00eb merrte muaj.<\/p>\n

    N\u00eb n\u00eb kursin online<\/a><\/noindex> Samsung \u201eRrjetet nervore dhe pamja kompjuterike\u201d, p\u00ebr nj\u00eb shembull, n\u00eb detyr\u00ebn finale Kaggle<\/a><\/noindex> me klasifikimin e pjata mbi t\u00eb pastra dhe t\u00eb pista, demonstrohet nj\u00eb qasje q\u00eb brenda 5 minutash ju jep n\u00eb dispozit\u00eb nj\u00eb rrjet nervor t\u00eb thell\u00eb, i cili \u00ebsht\u00eb n\u00eb gjendje t\u00eb dalloj\u00eb pjata t\u00eb pista nga t\u00eb pastra, i nd\u00ebrtuar sipas arkitektures s\u00eb p\u00ebrshkruar m\u00eb lart. Rrjeti fillestar nuk dinte se \u00e7far\u00eb ishin platat, ai thjesht m\u00ebsoi t\u00eb dallonte zogjt\u00eb nga qent\u00eb (shih ImageNet).<\/p>\n

    \"Grafiku
    \nBurimi:     kursin online<\/a><\/noindex> Samsung \u201eRrjetet nervore dhe pamja kompjuterike\u201d<\/i><\/p>\n

    P\u00ebr Transfer Learning \u00ebsht\u00eb e nevojshme t\u00eb dini se cilat qasje funksionojn\u00eb, cilat jan\u00eb arkitekturat themelore t\u00eb gatshme. N\u00eb p\u00ebrgjith\u00ebsi, kjo e p\u00ebrshpejton shum\u00eb paraqitjen e aplikacioneve praktike t\u00eb m\u00ebsimit t\u00eb makinerive.<\/p>\n

    1.2. Rrjetet Generative Adversarial (Generative Adversarial Networks, GAN)<\/h2>\n

    \nKy \u00ebsht\u00eb p\u00ebr raste kur na \u00ebsht\u00eb e v\u00ebshtir\u00eb t\u00eb formulojm\u00eb q\u00ebllimin e m\u00ebsimit. Sa m\u00eb af\u00ebr t\u00eb jet\u00eb detyra me jet\u00ebn reale, aq m\u00eb e qart\u00eb \u00ebsht\u00eb p\u00ebr ne (\"sillni tavolin\u00ebn\"), por aq m\u00eb e v\u00ebshtir\u00eb \u00ebsht\u00eb ta formulojm\u00eb si nj\u00eb detyr\u00eb teknike. GAN \u00ebsht\u00eb p\u00ebrpjekja p\u00ebr t\u00eb na shp\u00ebtuar nga kjo problematik\u00eb.<\/p>\n

    K\u00ebtu punojn\u00eb dy rrjete: nj\u00eb gjenerues (Generative), dhe tjetra diskiminator (Adversarial). Nj\u00eb rrjet m\u00ebson t\u00eb b\u00ebj\u00eb pun\u00eb t\u00eb dobishme (t\u00eb klasifikoj\u00eb imazhe, t\u00eb njoh\u00eb tinguj, t\u00eb vizatoj\u00eb karikatura). Nd\u00ebrsa rrjeti tjet\u00ebr m\u00ebson t\u00eb m\u00ebsoj\u00eb at\u00eb rrjet: ai ka mostra reale, dhe m\u00ebson t\u00eb gjej\u00eb nj\u00eb formul\u00eb komplekse q\u00eb nuk ishte e njohur m\u00eb par\u00eb p\u00ebr t\u00eb krahasuar prodhimet e pjes\u00ebs gjenerative t\u00eb rrjetit me objektet e bot\u00ebs reale (grupin m\u00ebsimor) p\u00ebr tipare t\u00eb thella t\u00eb r\u00ebnd\u00ebsishme: numrin e syve, af\u00ebrsin\u00eb me stilin e Miyazaki, sakt\u00ebsin\u00eb e shqiptes anglisht.<\/p>\n

    \"Grafiku
    \nNj\u00eb shembull t\u00eb rezultatit t\u00eb pun\u00ebs s\u00eb rrjetit p\u00ebr prodhimin e personazheve anime.     Burimi<\/a><\/noindex><\/i><\/p>\n

    Por sigurisht, \u00ebsht\u00eb e v\u00ebshtir\u00eb t\u00eb nd\u00ebrtohet nj\u00eb arkitektur\u00eb. Nuk mjafton thjesht t\u00eb hidhen neurone, ato duhet t\u00eb trajtohen. Dhe shpesh m\u00ebsohen p\u00ebr jav\u00eb t\u00eb t\u00ebra. Tema GAN \u00ebsht\u00eb nj\u00eb nga pyetjet ky\u00e7e q\u00eb koleg\u00ebt e mi n\u00eb Qendr\u00ebn e Inteligjenc\u00ebs Artificiale t\u00eb Samsung po e shqyrtojn\u00eb. P\u00ebr shembull, kjo zhvillimi<\/a><\/noindex>: p\u00ebrdorimi i rrjeteve gjeneruese p\u00ebr sintez\u00ebn e fotografive realiste t\u00eb njer\u00ebzve me pozita t\u00eb ndryshueshme \u2014 p\u00ebr shembull, p\u00ebr t\u00eb krijuar nj\u00eb dhom\u00eb provimi virtuale, ose p\u00ebr sintez\u00ebn e fytyr\u00ebs, q\u00eb mund t\u00eb leht\u00ebsoj\u00eb sasin\u00eb e informacionit q\u00eb duhet ruajtur ose transmision p\u00ebr t\u00eb siguruar video cil\u00ebsore, transmetim ose mbrojtje t\u00eb t\u00eb dh\u00ebnave personale.<\/p>\n

    \"Grafiku
    \n    Burimi<\/a><\/noindex><\/i><\/p>\n

    1.3. AI i Shpjeguesh\u00ebm (Explainable AI)<\/h2>\n

    \nN\u00eb disa raste t\u00eb rralla, p\u00ebrparimi n\u00eb arkitektur\u00ebn e thell\u00eb papritur afroi mund\u00ebsit\u00eb e rrjeteve nervore t\u00eb thella me ato njer\u00ebzore. Tani lufta po zhvillohet p\u00ebr t\u00eb zgjeruar k\u00ebt\u00eb grup detyrash. P\u00ebr shembull, nj\u00eb robot pastrues do t\u00eb mund t\u00eb dallonte leht\u00ebsisht nj\u00eb mace nga nj\u00eb qen gjat\u00eb nj\u00eb takimi ball\u00eb p\u00ebr ball\u00eb. Por n\u00eb shumic\u00ebn e situatave t\u00eb jet\u00ebs, ai do t\u00eb jet\u00eb i paaft\u00eb t\u00eb gjej\u00eb nj\u00eb mace q\u00eb po fle midis rrobave ose mobiljeve (n\u00eb t\u00eb v\u00ebrtet\u00eb, si\u00e7 jemi edhe ne n\u00eb shumic\u00ebn e rasteve\u2026).<\/p>\n

    Cili \u00ebsht\u00eb shkaku i suksesit t\u00eb rrjeteve nervore t\u00eb thella? Ato krijojn\u00eb p\u00ebrfaq\u00ebsimin e detyr\u00ebs, q\u00eb bazohet jo n\u00eb informacionin \"e dukshme me sy t\u00eb lir\u00eb\" (piksel\u00ebt e fotografis\u00eb, ndryshimet e z\u00ebrit...), por n\u00eb karakteristikat q\u00eb fitohen pas p\u00ebrpunimit t\u00eb k\u00ebsaj informazioni p\u00ebrmes disa qindra shtresave t\u00eb rrjetit nervor. Fatkeq\u00ebsisht, k\u00ebto lidhje gjithashtu mund t\u00eb jen\u00eb pa kuptim, t\u00eb kontradiktojn\u00eb nj\u00ebra-tjetr\u00ebn ose t\u00eb mbajn\u00eb shenja t\u00eb pap\u00ebrsosm\u00ebris\u00eb s\u00eb grupit fillestar t\u00eb t\u00eb dh\u00ebnave. P\u00ebr shembull, p\u00ebr at\u00eb q\u00eb mund t\u00eb ndodh\u00eb nga p\u00ebrdorimi i pav\u00ebmendsh\u00ebm t\u00eb AI n\u00eb rekrutimin e punonj\u00ebsve, ekziston nj\u00eb loj\u00eb e vog\u00ebl kompjuterike. Survival Of The Best Fit<\/a><\/noindex>.<\/p>\n

    \"Grafiku
    \nSistemi p\u00ebr etiketimin e imazheve e quajti nj\u00eb njeri q\u00eb po gatuan, grua, edhe pse n\u00eb fotografin\u00eb n\u00eb t\u00eb v\u00ebrtet\u00eb ishte nj\u00eb burr\u00eb (    Burimi<\/a><\/noindex>). Kjo v\u00ebrejt\u00ebm<\/a><\/noindex> n\u00eb Institutin e Virgjinis\u00eb.<\/i><\/p>\n

    P\u00ebr t\u00eb analizuar lidhjet komplekse dhe t\u00eb thella, t\u00eb cilat shpesh ne nuk mundemi t'i formulonim vet\u00eb, jan\u00eb t\u00eb nevojshme metodat e AI t\u00eb Shpjeguesh\u00ebm. Ato organizojn\u00eb karakteristikat e rrjeteve nervore t\u00eb thella n\u00eb m\u00ebnyr\u00eb q\u00eb pas trajnimit t\u00eb mund t\u00eb analizojm\u00eb p\u00ebrfaq\u00ebsimin e brendsh\u00ebm t\u00eb m\u00ebsuar nga rrjeti, e jo thjesht t\u00eb mb\u00ebshtetemi n\u00eb zgjidhjen e saj.<\/p>\n

    1.4. Analitika \/ AI n\u00eb Skaj (Edge Analytics \/ AI)<\/h2>\n

    \n\u00c7do gj\u00eb q\u00eb ka fjal\u00ebn Edge, do t\u00eb thot\u00eb n\u00eb m\u00ebnyr\u00eb t\u00eb drejtp\u00ebrdrejt\u00eb: transferimi i nj\u00eb pjese t\u00eb algoritmeve nga reja\/serveri n\u00eb nivel t\u00eb pajisjes p\u00ebrfundimtare\/gatishm\u00ebris\u00eb. Nj\u00eb algorit\u00ebm i till\u00eb do t\u00eb veproj\u00eb m\u00eb shpejt dhe nuk do t\u00eb ket\u00eb nevoj\u00eb t\u00eb lidhet me serverin qendror p\u00ebr t\u00eb funksionuar. N\u00ebse ju njohin abstraktin \"klienti i holl\u00eb\", k\u00ebtu e b\u00ebjm\u00eb k\u00ebt\u00eb klient pak m\u00eb t\u00eb trash\u00eb.
    \nKjo mund t\u00eb jet\u00eb e r\u00ebnd\u00ebsishme p\u00ebr Internetin e Gj\u00ebrave. P\u00ebr shembull, n\u00ebse nj\u00eb makin\u00eb \u00ebsht\u00eb tepruar dhe ka nevoj\u00eb p\u00ebr ftohje, ka kuptim t\u00eb d\u00ebrgohet nj\u00eb sinjal menj\u00ebher\u00eb, n\u00eb nivelin e fabrik\u00ebs, pa pritur q\u00eb t\u00eb dh\u00ebnat t\u00eb kalojn\u00eb n\u00eb re dhe pastaj te specialisti i turnit. Ose nj\u00eb tjet\u00ebr shembull: automjetet pa shofer mund t\u00eb kuptojn\u00eb situat\u00ebn e trafikut vet\u00eb, pa u drejtuar te serveri qendror.<\/p>\n

    \"Grafiku
    \n    Burimi<\/a><\/noindex><\/i><\/p>\n

    Ose nj\u00eb tjet\u00ebr shembull, pse kjo \u00ebsht\u00eb e r\u00ebnd\u00ebsishme nga k\u00ebndv\u00ebshtrimi i siguris\u00eb: kur ju shkruani tekste n\u00eb telefonin tuaj, ai e ruan fjal\u00ebt tipike p\u00ebr ju, q\u00eb m\u00eb pas tastiera e telefonit t'ju sugjeroj\u00eb ato n\u00eb m\u00ebnyr\u00eb t\u00eb p\u00ebrshtatshme \u2013 kjo quhet hyrje parashikuese e tekstit. D\u00ebrgimi i gjith\u00e7kaje q\u00eb shkruani n\u00eb qend\u00ebr t\u00eb t\u00eb dh\u00ebnave do t\u00eb ishte nj\u00eb shkelje e privat\u00ebsis\u00eb suaj dhe thjesht e pasigurt. Prandaj, trajnimi i tastier\u00ebs ndodh vet\u00ebm n\u00eb kuad\u00ebr t\u00eb pajisjes suaj.<\/p>\n

    1.5. Platforma e AI si sh\u00ebrbim (AI PaaS)<\/h2>\n

    \nPaaS \u2013 Platforma-si-sh\u00ebrbim \u2013 \u00ebsht\u00eb nj\u00eb model biznesi, ku ne kemi qasje n\u00eb nj\u00eb platform\u00eb t\u00eb integruar, p\u00ebrfshir\u00eb ruajtjen e saj cloud dhe procedurat e gatshme. N\u00eb k\u00ebt\u00eb m\u00ebnyr\u00eb, ne mund t\u00eb \u00e7lirohemi nga detyrat infrastrukturore, dhe t\u00eb p\u00ebrqendrohemi krejt\u00ebsisht n\u00eb prodhimin e di\u00e7kaje t\u00eb dobishme. Disa nga platforma PaaS p\u00ebr detyra AI jan\u00eb: IBM Cloud, Microsoft Azure, Amazon Machine Learning, Google AI Platform.<\/p>\n

    1.6. M\u00ebsimi adaptiv i makineris\u00eb (Adaptive ML)<\/h2>\n

    \n\u00c7far\u00eb ndodh n\u00ebse ne lejojm\u00eb inteligjenc\u00ebn artificiale t\u00eb p\u00ebrshtatet ... Ju do t\u00eb pyesni \u2013 si? ... A nuk p\u00ebrshtatet ajo tashm\u00eb me detyr\u00ebn? Problemi \u00ebsht\u00eb ky: \u00e7do detyr\u00eb t\u00eb till\u00eb ne e p\u00ebrpilojm\u00eb me kujdes para se t\u00eb nd\u00ebrtojm\u00eb algoritmin p\u00ebr zgjidhjen e saj nga inteligjenca artificiale. Ju do t\u00eb p\u00ebrgjigjen \u2013 duket se \u00ebsht\u00eb e mundur ta thjeshtojm\u00eb k\u00ebt\u00eb zinxhir.<\/p>\n

    M\u00ebsimi tradicional i makin\u00ebs punon sipas parimit t\u00eb sistemit t\u00eb hapur (open-loop): ju p\u00ebrgatitni t\u00eb dh\u00ebnat, shpikni nj\u00eb rrjet neuromotor (apo \u00e7far\u00ebdo tjet\u00ebr), e trajni, pastaj shikoni disa tregues dhe n\u00ebse ju p\u00eblqen, mund ta d\u00ebrgoni rrjetin neuromotor n\u00eb smartphone \u2013 p\u00ebr t\u00eb zgjidhur problemet e p\u00ebrdoruesve. Por n\u00eb aplikime ku ka shum\u00eb t\u00eb dh\u00ebna dhe karakteri i tyre ndryshon gradualisht, nevojiten metoda t\u00eb tjera. K\u00ebto sisteme, q\u00eb adaptohen dhe m\u00ebsojn\u00eb vet\u00eb, organizohen n\u00eb cikle t\u00eb mbyllura, t\u00eb vet\u00eb-m\u00ebsuara (closed-loop), dhe ato duhet t\u00eb funksionojn\u00eb pa nd\u00ebrprerje.<\/p>\n

    Aplikimet \u2013 mund t\u00eb jen\u00eb analiza e rrjedhave (Stream Analytics), mbi t\u00eb cilat shum\u00eb biznesmen\u00eb marrin vendime, ose menaxhimi adaptiv i produksionit. N\u00eb mas\u00ebn e aplikimeve t\u00eb sotme dhe duke marr\u00eb parasysh rreziqet m\u00eb t\u00eb kuptueshme p\u00ebr njer\u00ebzit, metodat q\u00eb p\u00ebrb\u00ebjn\u00eb zgjidhjen e k\u00ebtij problemi, t\u00eb gjitha k\u00ebto metoda grumbullohen n\u00ebn emrin e p\u00ebrgjithsh\u00ebm AI Adaptiv.<\/p>\n

    \"Grafiku
    \n    Burimi<\/a><\/noindex><\/i><\/p>\n

    Duke par\u00eb k\u00ebt\u00eb imazh, \u00ebsht\u00eb e v\u00ebshtir\u00eb t\u00eb largohemi nga ndjenja se futurist\u00ebt s\u2019kan\u00eb p\u00ebr t\u00eb ngr\u00ebn\u00eb \u2013 vet\u00ebm jepini m\u00ebsim nj\u00eb roboti se si t\u00eb marr\u00eb frym\u00eb...<\/p>\n

    Kompjuter\u00ebt dhe Komunikimet Postklasike (Postclassical Compute and Comms)<\/h1>\n

    <\/p>\n

    2.1. Rrjeti celular i gjenerat\u00ebs s\u00eb pest\u00eb (5G)<\/h2>\n

    \nKjo \u00ebsht\u00eb nj\u00eb tem\u00eb aq interesante, saq\u00eb menj\u00ebher\u00eb ju d\u00ebrgojm\u00eb te artikullin<\/a><\/noindex>. E k\u00ebtu \u00ebsht\u00eb nj\u00eb p\u00ebrmbledhje e shkurt\u00ebr. 5G, p\u00ebr shkak t\u00eb rritjes s\u00eb frekuenc\u00ebs s\u00eb transmetimit t\u00eb t\u00eb dh\u00ebnave, do ta b\u00ebj\u00eb shpejt\u00ebsin\u00eb e Internetit jasht\u00ebzakonisht t\u00eb shpejt\u00eb. Val\u00ebt e shkurtra kan\u00eb v\u00ebshtir\u00ebsi p\u00ebr t'u kaluar p\u00ebrmes pengesave, k\u00ebshtu q\u00eb struktura e rrjeteve do t\u00eb jet\u00eb krejt\u00ebsisht ndryshe: do t\u00eb nevojiten 500 her\u00eb m\u00eb shum\u00eb stacione baz\u00eb.<\/p>\n

    S\u00eb bashku me shpejt\u00ebsin\u00eb, do t\u00eb kemi shfaqje t\u00eb reja: lojra n\u00eb koh\u00eb reale me realitet t\u00eb shtuar, p\u00ebrfundimin e detyrave t\u00eb komplikuara (si kirurgjia) p\u00ebrmes tele-pranishm\u00ebris\u00eb, parandalimin e aksidenteve dhe situatave t\u00eb komplikuara n\u00eb rrug\u00eb p\u00ebrmes komunikimit midis automjeteve. Nga nj\u00eb perspektiv\u00eb m\u00eb praktike: p\u00ebrfundimisht do t\u00eb ndaloj\u00eb r\u00ebnia e Internetit celular gjat\u00eb ngjarjeve masive, si\u00e7 jan\u00eb ndeshjet n\u00eb stadium.<\/p>\n

    \"Grafiku
    \nBurimi i imazhit \u2013 Reuters, Niantic<\/i><\/p>\n

    2.2. Memoria e gjenerat\u00ebs s\u00eb ardhshme (Next-Generation Memory)<\/h2>\n

    \nK\u00ebtu b\u00ebhet fjal\u00eb p\u00ebr gjenerat\u00ebn e pest\u00eb t\u00eb memories operativ\u00eb \u2013 DDR5. Samsung ka njoftuar se deri n\u00eb fund t\u00eb vitit 2019 do t\u00eb shfaqen produkte mbi baz\u00ebn e DDR5. Pritet q\u00eb memoria e re t\u00eb jet\u00eb dy her\u00eb m\u00eb e shpejt\u00eb dhe dy her\u00eb m\u00eb kapacitet e madhe duke ruajtur form\u00ebn e saj, dometh\u00ebn\u00eb ne do t\u00eb mund t\u00eb marrim p\u00ebr kompjuterin ton\u00eb modulet e memories me kapacitet deri n\u00eb 32GB. N\u00eb t\u00eb ardhmen, kjo do t\u00eb jet\u00eb ve\u00e7an\u00ebrisht e r\u00ebnd\u00ebsishme p\u00ebr smartphone-t (memoria e re do t\u00eb jet\u00eb n\u00eb versionin me konsum t\u00eb ul\u00ebt t\u00eb energjis\u00eb) dhe p\u00ebr laptop\u00ebt (ku numri i slot-\u00ebve DIMM \u00ebsht\u00eb i kufizuar). Gjithashtu, m\u00ebsimi i makin\u00ebs k\u00ebrkon sasi t\u00eb m\u00ebdha memories operativ\u00eb.<\/p>\n

    2.3. Sistemet satelitore n\u00eb orbit\u00eb t\u00eb ul\u00ebt (Low-Earth-Orbit Satellite Systems)<\/h2>\n

    \nIdeja e z\u00ebvend\u00ebsimit t\u00eb satelit\u00ebve t\u00eb r\u00ebnd\u00eb, t\u00eb shtrenjt\u00eb dhe t\u00eb fuqish\u00ebm me nj\u00eb grumbull t\u00eb vog\u00ebl dhe t\u00eb lir\u00eb nuk \u00ebsht\u00eb aspak e re dhe ka lindur qysh n\u00eb vitet '90. P\u00ebr at\u00eb q\u00eb \u00abElon Musk s\u00eb shpejti do t'u jap\u00eb gjith\u00ebve Internet nga sateliti\u00bb<\/a><\/noindex> nuk ka d\u00ebgjuar tani, ka vet\u00ebm lenient. K\u00ebtu kompania m\u00eb e njohur \u00ebsht\u00eb Iridium, e cila shpalli falimentimin n\u00eb fund t\u00eb viteve '90, por u shp\u00ebtua nga Ministria e Mbrojtjes s\u00eb SHBA (mos e ngat\u00ebrro me iRidium \u2013 sistemin rus t\u00eb sht\u00ebpis\u00eb inteligjente). Projekti i Elon Musk (Starlink) nuk \u00ebsht\u00eb i vetmi \u2013 n\u00eb gar\u00ebn satelitore marrin pjes\u00eb Richard Branson (OneWeb \u2013 1440 satelit\u00eb t\u00eb supozuar), Boeing (3000 satelit\u00eb), Samsung (4600 satelit\u00eb), dhe tjer\u00eb.<\/p>\n

    Si \u00ebsht\u00eb situata n\u00eb k\u00ebt\u00eb fush\u00eb, si duket ekonomia \u2013 lexoni n\u00eb shqyrtimi<\/a><\/noindex>. Dhe ne presim provat e para t\u00eb k\u00ebtyre sistemeve nga p\u00ebrdoruesit e par\u00eb, t\u00eb cilat duhet t\u00eb ndodhin vitin tjet\u00ebr.<\/p>\n

    2.4. Printimi 3D n\u00eb nanoskal\u00eb (Nanoscale 3D Printing)<\/h2>\n

    \nPrintimi 3D, ndon\u00ebse nuk ka hyr\u00eb n\u00eb jet\u00eb t\u00eb \u00e7do personi (n\u00eb form\u00ebn q\u00eb premtonte fabrika plastike sht\u00ebpiake individuale), megjithat\u00eb ka dal\u00eb prej koh\u00ebsh nga ni\u0161a e teknologjive p\u00ebr geek. Mund t\u00eb gjykohet nga fakti se ekziston njohuri t\u00eb pakt\u00ebn p\u00ebr stilolapsat 3D nga \u00e7do nx\u00ebn\u00ebs, dhe shum\u00eb prej tyre \u00ebnd\u00ebrrojn\u00eb t\u00eb blejn\u00eb nj\u00eb kuti me gjurm\u00eb dhe ekstruzator p\u00ebr\u2026 \u00abthjesht k\u00ebshtu\u00bb (ose tashm\u00eb e kan\u00eb bler\u00eb).<\/p>\n

    Stereolitografia (printuesit 3D me lazer) lejojn\u00eb t\u00eb printohet me foton e ve\u00e7ant\u00eb: jan\u00eb duke u hetuar polimer\u00eb t\u00eb rinj, p\u00ebr ngurt\u00ebsimin e t\u00eb cil\u00ebve mjaftojn\u00eb dy fotona. Kjo do t\u00eb lejoj\u00eb q\u00eb n\u00eb kushte jo-laboratorike t\u00eb krijohen filtra t\u00eb rinj, mb\u00ebshtetje, springe, kapilare, lente dhe\u2026 propozimet tuaja n\u00eb komente! Dhe k\u00ebtu nuk \u00ebsht\u00eb larg nga fotopolimerizimi \u2013 vet\u00ebm kjo teknologji lejon t\u00eb \u00abprintohet\u00bb procesor\u00eb dhe skema llogarit\u00ebse. P\u00ebr m\u00eb tep\u00ebr, nuk ka kaluar vit q\u00eb ekziston teknologjia e printimit t\u00eb strukturave 3D t\u00eb grafenit 500-nm<\/a><\/noindex>, por pa zhvillim radikal.<\/p>\n

    \"Grafiku
    \n    Burimi<\/i><\/a><\/noindex><\/p>\n

    3. Sensorika dhe mobiliteti (Sensing and Mobility)<\/h1>\n

    <\/p>\n

    3.1. Automjetet autonome, niveli 4 dhe 5 (Autonomous Driving Level 4 & 5)<\/h2>\n

    \nP\u00ebr t\u00eb mos u ngat\u00ebrruar n\u00eb terminologji, \u00ebsht\u00eb e dobishme t\u00eb kuptohet se cilat nivele autonomi dallohen (marr\u00eb nga detajet artikullit<\/a><\/noindex>, t\u00eb cil\u00ebn e d\u00ebrgojm\u00eb p\u00ebr t\u00eb gjith\u00eb ata q\u00eb jan\u00eb t\u00eb interesuar):<\/p>\n

    Niveli 1: Kontrolli i kruizit: ndihm\u00eb p\u00ebr shoferin n\u00eb situata shum\u00eb t\u00eb kufizuara (p\u00ebr shembull \u2013 mbajtja e automjetit me nj\u00eb shpejt\u00ebsi t\u00eb caktuar pasi shoferi ka hequr k\u00ebmb\u00ebn nga pedali)
    \nNiveli 2: Ndihm\u00eb e kufizuar me drejtimin dhe frenimin. Shoferi duhet t\u00eb jet\u00eb i gatsh\u00ebm t\u00eb marr\u00eb kontrollin menj\u00ebher\u00eb. Duar n\u00eb timon, shikimi i orientuar n\u00eb rrug\u00eb. Kjo \u00ebsht\u00eb ajo q\u00eb tashm\u00eb ekziston n\u00eb Tesla dhe General Motors.
    \nNiveli 3: Shoferi nuk ka nevoj\u00eb t\u00eb ndjek\u00eb vazhdimisht rrug\u00ebn. Por ai duhet t\u00eb jet\u00eb n\u00eb gatishm\u00ebri dhe t\u00eb jet\u00eb i gatsh\u00ebm t\u00eb rimarr\u00eb kontrollin. Kjo \u00ebsht\u00eb ajo q\u00eb nuk ekziston ende n\u00eb automjetet e disponueshme p\u00ebr shitje. T\u00eb gjitha automjetet aktuale jan\u00eb n\u00eb nivelet 1-2.
    \nNiveli 4: Autopilot i v\u00ebrtet\u00eb, por me kufizime: vet\u00ebm udh\u00ebtime n\u00eb zona t\u00eb njohura, t\u00eb cilat jan\u00eb kartografuar me kujdes dhe n\u00eb p\u00ebrgjith\u00ebsi jan\u00eb t\u00eb njohura p\u00ebr sistemin, dhe n\u00eb kushte t\u00eb caktuara: p\u00ebr shembull, n\u00eb munges\u00eb t\u00eb bor\u00ebs. Ka prototipa nga Waymo dhe General Motors, dhe ata planifikojn\u00eb t\u2019i lancojn\u00eb n\u00eb disa qytete dhe t\u2019i testojn\u00eb n\u00eb nj\u00eb mjedis real. \u00abYandex\u00bb ka zona testimi p\u00ebr taksi pa shofer n\u00eb Skolkovo dhe Innopolis: udh\u00ebtimi ndodh n\u00ebn mbik\u00ebqyrjen e nj\u00eb inxhinieri q\u00eb ndodhet n\u00eb vendin e pasagjerit; deri n\u00eb fund t\u00eb vitit, kompania planifikon t\u00eb zgjedh\u00eb flot\u00ebn n\u00eb 100 automjete pa shofer.
    \nNiveli 5: Drejtim t\u00eb plot\u00eb automatik, z\u00ebvend\u00ebsim t\u00eb plot\u00eb t\u00eb shoferit t\u00eb gjall\u00eb. Sistemet e tilla nuk jan\u00eb ekzistuese, dhe \u00ebsht\u00eb e dyshimt\u00eb se do t\u00eb shfaqen n\u00eb vitet e ardhshme.<\/p>\n

    Sa realiste \u00ebsht\u00eb t\u00eb shohim gjith\u00e7ka k\u00ebt\u00eb n\u00eb nj\u00eb t\u00eb ardhme t\u00eb af\u00ebrt? K\u00ebtu do t\u00eb doja t\u2019i rekomandoja lexuesit artikullin \u00abPse lan\u00e7imi i robotaksive deri n\u00eb vitin 2020, si\u00e7 premton Tesla, \u00ebsht\u00eb i pamundur\u00bb<\/a><\/noindex>. Kjo \u00ebsht\u00eb pjes\u00ebrisht p\u00ebr shkak t\u00eb munges\u00ebs s\u00eb lidhjes 5G: shpejt\u00ebsit\u00eb e disponueshme 4G nuk jan\u00eb t\u00eb mjaftueshme. Pjes\u00ebrisht p\u00ebr shkak t\u00eb kostos shum\u00eb t\u00eb lart\u00eb t\u00eb automjeteve autonome: ato p\u00ebr momentin nuk jan\u00eb rentab\u00ebl, dhe model biznesi \u00ebsht\u00eb i paqart\u00eb. N\u00eb nj\u00eb fjal\u00eb, \u00abgjith\u00e7ka \u00ebsht\u00eb e komplikuar\u00bb, dhe nuk \u00ebsht\u00eb rast\u00ebsi q\u00eb Gartner shkruan se parashikimi i p\u00ebrdorimit masiv t\u00eb Nivelit 4 dhe 5 \u00ebsht\u00eb jo m\u00eb her\u00ebt se pas 10 vitesh.<\/p>\n

    3.2. Kamerat me shikim 3D (3D Sensing Cameras)<\/h2>\n

    \nTrafiku nga kontrolluesi i loj\u00ebrave Microsoft Kinect b\u00ebri zhurm\u00eb tet\u00eb vjet m\u00eb par\u00eb, duke ofruar nj\u00eb zgjidhje t\u00eb p\u00ebrballueshme dhe relativisht t\u00eb lir\u00eb p\u00ebr shikimin 3D. Q\u00eb nga ajo koh\u00eb, loj\u00ebrat sportive dhe vall\u00ebzuese me Kinect kan\u00eb pasur nj\u00eb ngritje dhe r\u00ebnien e tyre t\u00eb shpejt\u00eb, por kamerat 3D jan\u00eb p\u00ebrdorur n\u00eb robot\u00ebt industrial\u00eb, automjetet pa shofer, telefona mobil\u00eb p\u00ebr identifikim p\u00ebrmes fytyr\u00ebs. T\u00eb dh\u00ebnat teknologjike jan\u00eb b\u00ebr\u00eb m\u00eb t\u00eb lira, kompakte dhe m\u00eb t\u00eb aksesueshme. <\/p>\n

    \"Grafiku
    \nN\u00eb telefonin Samsung S10 ka nj\u00eb kamer\u00eb me koh\u00ebn e fluturimit (Time-of-Flight), e cila mat distanc\u00ebn deri te objekti \u2013 p\u00ebr t\u00eb leht\u00ebsuar fokusimin.     Burimi<\/a><\/noindex><\/i><\/p>\n

    N\u00ebse jeni t\u00eb interesuar p\u00ebr k\u00ebt\u00eb tem\u00eb, ju rekomandojm\u00eb nj\u00eb p\u00ebrmbledhje t\u00eb shk\u00eblqyer t\u00eb kamerave t\u00eb thell\u00ebsis\u00eb: pjesa 1<\/a><\/noindex>, pjesa 2<\/a><\/noindex>.<\/p>\n

    3.3. Dron\u00ebt p\u00ebr d\u00ebrgimin e ngarkesave t\u00eb vogla (Light Cargo Delivery Drones)<\/h2>\n

    \nK\u00ebt\u00eb vit, Amazon b\u00ebri zhurm\u00eb kur tregoi n\u00eb nj\u00eb ekspozit\u00eb dronin e ri fluturues, q\u00eb mund t\u00eb transportoj\u00eb ngarkesa t\u00eb vogla deri n\u00eb 2 kg pesh\u00eb. P\u00ebr qytetin, me trafikun e tij, kjo duket si zgjidhja ideale. Le t\u00eb shohim si do t\u00eb performojn\u00eb k\u00ebto dron\u00eb n\u00eb t\u00eb ardhmen e af\u00ebrt. Besoj se \u00ebsht\u00eb e r\u00ebnd\u00ebsishme t\u00eb p\u00ebrfshihet nj\u00eb skepticiz\u00ebm i kujdessh\u00ebm: ka shum\u00eb probleme, duke filluar nga mund\u00ebsia e vjedhjes s\u00eb leht\u00eb t\u00eb dronit, e deri te kufizimet ligjore mbi BPLAs. Amazon Prime Air ekziston p\u00ebr gjasht\u00eb vjet, por ende \u00ebsht\u00eb n\u00eb faz\u00ebn e testimit.<\/p>\n

    \"Grafiku
    \nDroni i ri Amazon, i paraqitur k\u00ebt\u00eb pranver\u00eb. Ka di\u00e7ka prej \u2018Yll\u00ebsive t\u00eb Luft\u00ebs\u2019 n\u00eb t\u00eb.     Burimi<\/a><\/noindex><\/i><\/p>\n

    P\u00ebrve\u00e7 Amazonit, ka edhe lojtar\u00eb t\u00eb tjer\u00eb n\u00eb k\u00ebt\u00eb treg (ka nj\u00eb p\u00ebrmbledhje t\u00eb detajuar rishikimin ton\u00eb<\/a><\/noindex>), por asnj\u00eb produkt n\u00eb dispozicion: gjith\u00e7ka \u00ebsht\u00eb n\u00eb faz\u00ebn e testimit dhe aktiviteteve marketingu. Ve\u00e7an\u00ebrisht interesante p\u00ebr t\u00eb theksuar jan\u00eb dron\u00ebt e specializuar mjek\u00ebsor\u00eb projektesh<\/a><\/noindex> n\u00eb Afrik\u00eb: d\u00ebrgimi i gjakut t\u00eb dhuruar n\u00eb Gana (14,000 d\u00ebrgesa, kompania Zipline) dhe Ruanda (kompania Matternet).<\/p>\n

    3.4. Transporti autonom fluturues (Flying Autonomous Vehicles)<\/h2>\n

    \nK\u00ebtu \u00ebsht\u00eb e v\u00ebshtir\u00eb t\u00eb thuash di\u00e7ka t\u00eb sakt\u00eb. Sipas Gartner, kjo do t\u00eb shfaqet m\u00eb her\u00ebt se pas 10 vitesh. N\u00eb p\u00ebrgjith\u00ebsi, k\u00ebtu jan\u00eb t\u00eb nj\u00ebjtat probleme si me automjetet pa shofer, por ato marrin nj\u00eb dimension t\u00eb ri \u2014 vertikal. Cilat jan\u00eb ambiciet p\u00ebr nd\u00ebrtimin e taksive fluturuese deklarojn\u00eb Porsche, Boeing, Uber.<\/p>\n

    3.5. N\u00eb re e realitetit t\u00eb pasur (AR Cloud)<\/h2>\n

    \nNj\u00eb kopje digjitale e p\u00ebrhershme e bot\u00ebs reale, q\u00eb lejon krijimin e nj\u00eb shtrese t\u00eb re realiteti, t\u00eb p\u00ebrbashk\u00ebt p\u00ebr t\u00eb gjith\u00eb p\u00ebrdoruesit. N\u00ebse flasim me nj\u00eb gjuh\u00eb m\u00eb teknike, b\u00ebhet fjal\u00eb p\u00ebr krijimin e nj\u00eb platforme t\u00eb hapur cloud, n\u00eb t\u00eb cil\u00ebn zhvilluesit mund t\u00eb integrojn\u00eb aplikacionet e tyre AR. Modeli i monetizimit \u00ebsht\u00eb i qart\u00eb, \u00ebsht\u00eb nj\u00eb analog i Steam. Ideja \u00ebsht\u00eb aq e thell\u00eb, sa tashm\u00eb disa besojn\u00eb se AR pa cloud \u00ebsht\u00eb thjesht e pavler\u00eb.<\/p>\n

    Si mundi t\u00eb duket n\u00eb t\u00eb ardhmen, \u00ebsht\u00eb ilustruar n\u00eb nj\u00eb video t\u00eb shkurt\u00ebr. Duket si nj\u00eb tjet\u00ebr episod i 'Mirrors' t\u00eb Zeza:
    \n

    \r\n \"Luaj\r\n
    <\/div>\r\n <\/div><\/center><\/p>\n

    Mund t\u00eb lexoni gjithashtu n\u00eb artikullin e rishikimit.<\/a><\/noindex><\/p>\n

    4. Njeriu 'i Shtuar' (Augmented Human)<\/h1>\n

    <\/p>\n

    4.1. AI p\u00ebr emocionet (Emotion AI)<\/h2>\n

    \nSi t\u00eb masim, simulojm\u00eb dhe reagojm\u00eb ndaj emocioneve njer\u00ebzore? Disa nga klient\u00ebt k\u00ebtu jan\u00eb kompanit\u00eb q\u00eb prodhojn\u00eb asistent\u00eb me z\u00eb si Amazon Alexa. Ata do t\u00eb mund t\u00eb integrosen n\u00eb sht\u00ebpit\u00eb tona n\u00ebse m\u00ebsojn\u00eb t\u00eb identifikojn\u00eb disponimin: t\u00eb kuptojn\u00eb shkakun e pak\u00ebnaq\u00ebsis\u00eb s\u00eb p\u00ebrdoruesit dhe t\u00eb provojn\u00eb t\u00eb korrigjojn\u00eb situat\u00ebn. N\u00eb t\u00eb v\u00ebrtet\u00eb, n\u00eb kontekst ka shum\u00eb m\u00eb tep\u00ebr informacion sesa n\u00eb mesazhin vet\u00eb. Dhe konteksti \u2013 \u00ebsht\u00eb shprehja e fytyr\u00ebs, intonacioni dhe sjellja jo-verbale.<\/p>\n

    Nga aplikimet e tjera praktike: analiza e emocioneve gjat\u00eb intervistave p\u00ebr pun\u00eb (n\u00ebp\u00ebrmjet video-intervistave), vler\u00ebsimi i reagimit ndaj reklama ose p\u00ebrmbajtjeve t\u00eb tjera video (buz\u00ebqeshje, qeshje), ndihma n\u00eb m\u00ebsim (p\u00ebr shembull, p\u00ebr praktikat e vetme n\u00eb artin e fjalimeve publike). <\/p>\n

    P\u00ebr k\u00ebt\u00eb tem\u00eb \u00ebsht\u00eb e v\u00ebshtir\u00eb t\u00eb shprehim mendimin m\u00eb mir\u00eb se autori i filmid t\u00eb shkurt\u00ebr 6-minut\u00ebsh Stealing Ur Feeling<\/a><\/noindex>. N\u00eb nj\u00eb video t\u00eb b\u00ebr\u00eb me shkatht\u00ebsi dhe stil, shikohet se si mund t\u00eb mat\u00ebsh emocionet tona p\u00ebr q\u00ebllime marketingu, dhe nga reagimet momentale t\u00eb fytyr\u00ebs tuaj mund t\u00eb kuptoni: n\u00ebse e doni pic\u00ebn, qen\u00ebt, Kanye West, dhe madje cila \u00ebsht\u00eb niveli juaj i t\u00eb ardhurave dhe IQ i p\u00ebrllogaritur. Kur t\u00eb kaloni n\u00eb faqen e filmit p\u00ebr lidhjen e sip\u00ebrme, b\u00ebheni pjes\u00ebmarr\u00ebs n\u00eb nj\u00eb video interaktive duke p\u00ebrdorur kamer\u00ebn e integruar t\u00eb kompjuterit tuaj. Filmi \u00ebsht\u00eb shfaqur n\u00eb disa festivale filmi.<\/p>\n

    \"Grafiku
    \n    Burimi<\/a><\/noindex><\/i><\/p>\n

    Ka madje nj\u00eb studim interesante: si t\u00eb njihni ironin\u00eb n\u00eb tekst. Ata mor\u00ebn twit\u00ebt me hashtag #sarcasm dhe krijuan nj\u00eb most\u00ebr m\u00ebsimore nga 25,000 twit\u00eb me ironin\u00eb dhe 100,000 twit\u00eb t\u00eb zakonsh\u00ebm p\u00ebr \u00e7do gj\u00eb. Ata p\u00ebrdor\u00ebn bibliotek\u00ebn TensorFlow, e trajnuan sistemin, ja rezultati:<\/p>\n

    \"Grafiku
    \n    Burimi<\/a><\/noindex><\/i><\/p>\n

    Prandaj tani, n\u00ebse nuk jeni t\u00eb sigurt p\u00ebr nj\u00eb koleg apo mik \u2013 a ju tha di\u00e7ka seriozisht apo me ironi, \u2013 mund t\u00eb p\u00ebrdorni tashm\u00eb rrjetin neural t\u00eb trajnuar<\/a><\/noindex>! <\/p>\n

    4.2. Inteligjenca e Shtuar (Augmented Intelligence)<\/h2>\n

    \nAutomatizimi i pun\u00ebs intelektuale p\u00ebrmes metodave t\u00eb m\u00ebsimit t\u00eb makin\u00ebs. Mund t\u00eb duket si di\u00e7ka e zakonshme? Por k\u00ebtu r\u00ebnd\u00ebsi ka formulimi, p\u00ebr m\u00eb tep\u00ebr, q\u00eb p\u00ebrputhet me akronimin e Inteligjenc\u00ebs Artificiale. Kjo na referon n\u00eb polemikat mbi 'inteligjenc\u00ebn e fuqishme' dhe 'inteligjenc\u00ebn e dob\u00ebt'.
    \nInteligjenca e fuqishme \u2013 \u00ebsht\u00eb ajo inteligjenc\u00eb artificiale nga filmat fantastiko-shkencor\u00eb, e cila \u00ebsht\u00eb plot\u00ebsisht ekuivalente me mendimin njer\u00ebzor dhe e kupton veten si nj\u00eb individ. Kjo ende nuk ekziston dhe nuk dihet n\u00ebse do t\u00eb ekzistoj\u00eb ndonj\u00ebher\u00eb.<\/p>\n

    Inteligjenca e dob\u00ebt \u2013 nuk \u00ebsht\u00eb nj\u00eb individ autonom, por asistent i njeriut. Ai nuk pretendon t\u00eb ket\u00eb nj\u00eb mendim si njeriu, thjesht di t\u00eb zgjidh\u00eb probleme informacioni, p\u00ebr shembull, t\u00eb identifikoj\u00eb \u00e7far\u00eb \u00ebsht\u00eb n\u00eb nj\u00eb foto ose t\u00eb p\u00ebrkthej\u00eb nj\u00eb tekst.<\/p>\n

    \"Grafiku
    \n    Burimi <\/a><\/noindex><\/i><\/p>\n

    N\u00eb k\u00ebt\u00eb kuptim, Inteligjenca e Shtuar \u00ebsht\u00eb nj\u00eb 'inteligjenc\u00eb e dob\u00ebt' e past\u00ebr, dhe formulimi duket i suksessh\u00ebm, pasi nuk krijon konfuzion dhe joshje p\u00ebr t\u00eb par\u00eb k\u00ebtu at\u00eb 'inteligjenc\u00ebn e fuqishme', p\u00ebr t\u00eb cil\u00ebn t\u00eb gjith\u00eb \u00ebnd\u00ebrrojn\u00eb (apo frik\u00ebsohen, n\u00ebse kujtojm\u00eb diskutimet e shumta mbi 'rebellionin e makinave'). Duke p\u00ebrdorur shprehjen Inteligjenca e Shtuar, ne menj\u00ebher\u00eb si t\u00eb jemi personazhe t\u00eb nj\u00eb filmi tjet\u00ebr: nga fantastika shkencore (si 'Un\u00eb, robot' i Asimovit) kalojm\u00eb n\u00eb cyberpunk ('aumentimet' n\u00eb k\u00ebt\u00eb zhan\u00ebr quhen t\u00eb gjitha implantat q\u00eb zgjerjn\u00eb mund\u00ebsit\u00eb e njeriut).<\/p>\n

    Si than\u00eb<\/a><\/noindex> Erik Brynjolfsson dhe Andrew McAfee: 'N\u00eb 10 vitet e ardhshme do t\u00eb ndodhin k\u00ebto. Jo inteligjenca artificiale do t\u00eb z\u00ebvend\u00ebsoj\u00eb menaxher\u00ebt, por ata menaxher\u00eb q\u00eb p\u00ebrdorin inteligjenc\u00ebn artificiale do t\u00eb z\u00ebvend\u00ebsojn\u00eb ata q\u00eb nuk e kan\u00eb b\u00ebr\u00eb ende'<\/p>\n

    Shembuj:<\/p>\n