Kompjuterat mund t\u00eb b\u00ebjn\u00eb shum\u00eb gj\u00ebra n\u00eb vendin ton\u00eb, si\u00e7 e ilustron shpikja e re e Google. Grupe shkenc\u00ebtar\u00ebsh t\u00eb Google dhe Stanfort kan\u00eb krijuar nj\u00eb program t\u00eb aft\u00eb, jo p\u00ebr t\u00eb njohur nj\u00eb objekt n\u00eb fotot tuaja, por p\u00ebr t\u00eb identifikuar gjith\u00eb sken\u00ebn dhe em\u00ebrtuar imazhin n\u00eb anglisht, sipas nj\u00eb shkrimi t\u00eb botuar n\u00eb Nju Jork Tajms. Ky sistem i ri n\u00ebntitrash, i pag\u00ebzuar me emrin Neural Legend (NIC), bazohet te teknikat e t\u00eb parit nga kompjuteri dhe trajtimit t\u00eb gjuh\u00ebs. K\u00ebrkuesit e Google thon\u00eb se jan\u00eb frym\u00ebzuar nga p\u00ebrparimi i p\u00ebrkthimit automatik: \u00ab…nj\u00eb rrjet neuronesh transformon nj\u00eb fraz\u00eb n\u00eb fr\u00ebngjisht n\u00eb nj\u00eb p\u00ebrfaq\u00ebsim vektorial dhe nj\u00eb tjet\u00ebr rrjet e p\u00ebrdor k\u00ebt\u00eb p\u00ebrfaq\u00ebsim p\u00ebr t\u00eb krijuar nj\u00eb fraz\u00eb n\u00eb gjermanisht\u00bb. N\u00eb vend t\u00eb rrjetit gjuh\u00ebsor, studiesit p\u00ebrdor\u00ebn nj\u00eb rrjet tjet\u00ebr q\u00eb sh\u00ebrben p\u00ebr t\u00eb klasifikuar objektet n\u00eb imazhe. Nj\u00eb rrjet, pra, kodon imazhin, nd\u00ebrkoh\u00eb q\u00eb tjetri krijon nj\u00eb fraz\u00eb shpjegimi. Sistemi \u00ebsht\u00eb eksperimentuar me shum\u00eb imazhe t\u00eb ndryshme. Cil\u00ebsia e em\u00ebrtimit t\u00eb dh\u00ebn\u00eb \u00ebsht\u00eb vler\u00ebsuar si e k\u00ebnaqshme nga nj\u00eb algorit\u00ebm q\u00eb vler\u00ebson cil\u00ebsin\u00eb e p\u00ebrkthimeve mes gjuh\u00ebve. Kjo inteligjenc\u00eb artificiale lejon para se gjithash t\u00eb p\u00ebrmir\u00ebsohet k\u00ebrkimi i figurave n\u00eb internet. \u00abK\u00ebshtu, mund t\u00eb katalogohen m\u00eb mir\u00eb dhe t\u00eb k\u00ebrkohen miliarda imazhe e video t\u00eb disponueshme n\u00eb internet, p\u00ebrshkrimi dhe arkivimi i t\u00eb cilave shpesh \u00ebsht\u00eb i keq. P\u00ebr momentin, sitet e k\u00ebrkimit si Google mb\u00ebshteten shum\u00eb te gjuha e shkruar q\u00eb shoq\u00ebron nj\u00eb video, ose nj\u00eb imazh, p\u00ebr t\u00eb p\u00ebrcaktuar at\u00eb q\u00eb ato p\u00ebrmbajn\u00eb\u00bb. Po programi i ri do t\u00eb jet\u00eb m\u00eb i p\u00ebrparuar. \u00abP\u00ebr shembull, nd\u00ebrkoh\u00eb q\u00eb nj\u00eb sistem njohjeje objekti klasik mund t\u00eb jet\u00eb i aft\u00eb t\u00eb njoh\u00eb nj\u00eb mace dhe nj\u00eb peshk t\u00eb kuq n\u00eb nj\u00eb imazh, sistemi i ri mund t\u00eb p\u00ebrcaktoj\u00eb q\u00eb skena p\u00ebrfaq\u00ebson n\u00eb fakt nj\u00eb mace q\u00eb kap nj\u00eb peshk, n\u00eb nj\u00eb akuarium. I bashkuar me nj\u00eb baz\u00eb njohurish q\u00eb p\u00ebrfshin marr\u00ebdh\u00ebnien grabitqar-pre mes dy kafsh\u00ebve, i nj\u00ebjti sistem mund t\u00eb jet\u00eb i aft\u00eb t\u00eb parashikoj\u00eb se peshku i kuq s\u00eb shpejti do t\u00eb p\u00ebrfundoj\u00eb n\u00eb kthetrat e maces dhe do t\u00eb hahet prej saj\u00bb. N\u00eb gazet\u00eb n\u00ebnvizohet fakti se kjo teknologji, me kalimin e koh\u00ebs, mund t\u00eb ndihmoj\u00eb t\u00eb verbrit dhe ata me shikim t\u00eb kufizuar, p\u00ebr t\u00eb kuptuar p\u00ebrmbajtjen e nj\u00eb imazhi. Robot\u00ebt e integruar n\u00eb vetura mund t\u00eb marrin vendime m\u00eb t\u00eb mira n\u00eb funksion t\u00eb kontekstit, por kjo mund t\u00eb ket\u00eb pasoja edhe n\u00eb efikasitetin e survejimit: \u00abK\u00ebta 15 vjet\u00ebt e fundit, kamerat e v\u00ebzhgimit jan\u00eb vendosur n\u00eb nj\u00eb num\u00ebr t\u00eb madhe hap\u00ebsirash publike dhe private. N\u00eb t\u00eb ardhmen, programi q\u00eb i v\u00eb n\u00eb pun\u00eb k\u00ebto kamera nuk do t\u00eb jet\u00eb i aft\u00eb vet\u00ebm t\u00eb identifikoj\u00eb njer\u00ebz t\u00eb caktuar, n\u00ebp\u00ebrmjet njohjes s\u00eb fytyr\u00ebs, thon\u00eb ekspert\u00ebt, por edhe disa lloje sjelljeje, mbase edhe t\u00eb lajm\u00ebroj\u00eb n\u00eb m\u00ebnyr\u00eb automatike autoritetet\u00bb. <\/p>\n","protected":false},"excerpt":{"rendered":"
Kompjuterat mund t\u00eb b\u00ebjn\u00eb shum\u00eb gj\u00ebra n\u00eb vendin ton\u00eb, si\u00e7 e ilustron shpikja e re e Google. Grupe shkenc\u00ebtar\u00ebsh t\u00eb Google dhe Stanfort kan\u00eb krijuar nj\u00eb program t\u00eb aft\u00eb, jo p\u00ebr t\u00eb njohur nj\u00eb objekt n\u00eb fotot tuaja, por p\u00ebr t\u00eb identifikuar gjith\u00eb sken\u00ebn dhe em\u00ebrtuar imazhin n\u00eb anglisht, sipas nj\u00eb shkrimi t\u00eb botuar n\u00eb […]<\/p>\n","protected":false},"author":5,"featured_media":8332,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[46],"tags":[],"class_list":["post-8331","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-enciklopedi"],"_links":{"self":[{"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/posts\/8331","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/comments?post=8331"}],"version-history":[{"count":0,"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/posts\/8331\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/media\/8332"}],"wp:attachment":[{"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/media?parent=8331"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/categories?post=8331"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/intervista.al\/web\/wp-json\/wp\/v2\/tags?post=8331"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}