{"id":17072,"date":"2025-03-21T21:24:00","date_gmt":"2025-03-21T20:24:00","guid":{"rendered":"https:\/\/letrat.eu\/?p=17072"},"modified":"2026-04-29T20:52:23","modified_gmt":"2026-04-29T19:52:23","slug":"kurre-mos-u-shtir-gjoja-fotograf-me-iphone","status":"publish","type":"post","link":"https:\/\/letrat.eu\/?p=17072","title":{"rendered":"Fjala e fundit e IA p\u00ebr foto-realizmin #1-3"},"content":{"rendered":"

(sg – sky division, 9 prill 2025 – ca portrete, dhe koment)<\/em><\/p>\n

(…n\u00ebse t’na ishte IA ndonj\u00eb person i men\u00e7ur, ndonj\u00eb shkenc\u00ebtar, vitin e kaluar do i takonte Nobel i mir\u00ebfillt\u00eb<\/strong> : ) <\/p>\n

Them se disa segmente t\u00eb shkenc\u00ebs e t\u00eb teknologjis\u00eb v\u00ebrtet p\u00ebr t’u mahnitur sa shpejt zhvillohen dhe avancohen. Ka ca tjerash, ca shkencash thuase u ngjajn\u00eb “breshkave”, u duhet 1 shekull p\u00ebr ta hedhur nga 1 hap t\u00eb vog\u00ebl aty-k\u00ebtu : ) Mir\u00ebpo s’\u00ebsht\u00eb me t\u00eb gjitha nj\u00ebsoj, nganj\u00ebher\u00eb gj\u00ebrat l\u00ebvizin aq shpejt, pothuaj sikur na merret dhe mendja nga shpejt\u00ebsia. <\/p>\n

Inteligjenca Artificiale (IA) ve\u00e7 e meriton mir\u00ebnjohjen e bot\u00ebs shkencore p\u00ebr ndihm\u00ebn n\u00eb fush\u00ebn e kimis\u00eb, sidoqoft\u00eb po i shtoj n\u00eb fund dhe pak fjal\u00eb rreth shkenc\u00ebs prapa IA-imazheve – shkencat ashtu-k\u00ebshtu t\u00eb lidhura.
\nSi\u00e7 e dim\u00eb, u ndan\u00eb madje dhe Nobela p\u00ebr zbulimet kryer me ndihm\u00ebn e IA (vitin e kaluar, p\u00ebr AlphaFold2), po t’mos ishte IA do duheshin ndoshta dhe dekada tjera. Parashikimi i strukturave proteinike tredimensionale ishte sfidues p\u00ebr dekada t\u00eb t\u00ebra, ve\u00e7 pat\u00ebn kaluar dekada duke u marr\u00eb me t\u00eb nj\u00ebjt\u00ebn problematik\u00eb (rreth struktur\u00ebs s\u00eb proteinave) por arritjet pat\u00ebn qen\u00eb “minimale”, krahasuar me hovin pas ndihm\u00ebs s\u00eb IA. <\/p>\n

Ky kontribut specifik i IA ishte m\u00eb i madhi dhe m\u00eb i r\u00ebnd\u00ebsishmi, kontribut direkt i Inteligjenc\u00ebs Artificiale p\u00ebr avancimin e njohurive n\u00eb fush\u00ebn e gjer\u00eb shkencore si\u00e7 \u00ebsht\u00eb Kimia. Dhe dihet sa t\u00eb r\u00ebnd\u00ebsishme jan\u00eb proteinat, ato i kontrollojn\u00eb dhe i drejtojn\u00eb t\u00eb gjitha reaksionet kimike, t\u00eb cilat s\u00eb bashku jan\u00eb vet\u00eb baza e jet\u00ebs, ashtu thuhet sepse pa to s’mund t’ket\u00eb as ekzistenc\u00eb as jet\u00eb. Ku ka m\u00eb e r\u00ebnd\u00ebsishme se kaq : )
\nDhe shih pothuaj si “ironi”… pik\u00ebrisht n\u00eb k\u00ebt\u00eb drejtim t\u00eb shkencave rreth JET\u00cbS do vinte kontributi jasht\u00ebzakonisht i madh i Inteligjenc\u00ebs Artificiale. <\/p>\n

Vitin e kaluar, 2024, \u00e7mimi Nobel p\u00ebr Kimi u nda nga 3 kontributor\u00eb t\u00eb shkenc\u00ebs, gjysma e Nobelit i takoi laureatit David Baker “p\u00ebr dizajnimin e kompjuterizuar t\u00eb proteinave”, nd\u00ebrsa gjysm\u00ebn tjet\u00ebr e ndan\u00eb Demis Hassabis dhe John Jumper “p\u00ebr parashikimin e struktur\u00ebs s\u00eb proteinave”. Dmth. Demis Hassabis u nderua me 1\/4 e Nobelit, dhe John Jumper me 1\/4 tjet\u00ebr.<\/p>\n

N\u00eb v. 2020, Demis Hassabis dhe John Jumper e prezantuan nj\u00eb model t\u00eb IA t\u00eb quajtur AlphaFold2. Me ndihm\u00ebn e k\u00ebtij modeli ata do ishin n\u00eb gjendje t\u00eb parashikojn\u00eb struktur\u00ebn e pothuajse t\u00eb gjitha proteinave t\u00eb njohura. AlphaFold2 (AF2) \u00ebsht\u00eb zhvilluar nga DeepMind (e Google, n\u00ebn udh\u00ebheqjen e D. Hassabis), dhe pas dekadave t\u00eb “ngadalshme” me proteinat tashm\u00eb dihen pothuaj t\u00eb gjitha strukturat e proteinave t\u00eb njohura deri sot. <\/p>\n

Dhe kjo \u00ebsht\u00eb di\u00e7 jasht\u00ebzakonisht e madhe, duhet cekur se r\u00ebnd\u00ebsia e k\u00ebsaj ndihme t\u00eb IA \u00ebsht\u00eb e jasht\u00ebzakonshme, shkenc\u00ebtar\u00ebt thon\u00eb se parashikimi i sakt\u00eb i strukturave t\u00eb proteinave mund ta p\u00ebrshpejtoj\u00eb dramatikisht kuptimin rreth mekanizmave biologjik\u00eb, dhe k\u00ebshtu t\u00eb siguroj\u00eb nj\u00eb baz\u00eb t\u00eb fort\u00eb p\u00ebr zhvillimin e farmakologjis\u00eb, p\u00ebr zhvillimin e sigurt t\u00eb ila\u00e7eve dhe t\u00eb barnave t\u00eb ndryshme. Dihet sa t\u00eb r\u00ebnd\u00ebsishme jan\u00eb ila\u00e7et… no comment, dhe dihet kujt duhet “hequr kapel\u00ebn” (IA-s) prap\u00eb no comment : )<\/p>\n

***
\nFjala e fundit e IA p\u00ebr foto-realizmin – #3<\/strong><\/p>\n

Kur \u00ebsht\u00eb fjala p\u00ebr “shpejt\u00ebsin\u00eb” e zhvillimeve, ngjitur s\u00eb bashku me disa portrete-eksperimente, po i shtoj dhe 2-3 fjal\u00eb sa p\u00ebr ta shpalosur mahnitjen rreth dinamizmit zhvillimor t\u00eb IA-s
\nN\u00eb Shtator 2024, nj\u00eb grumbull autor\u00ebsh rrotull platform\u00ebs “Playground Research” e publikuan dokumentin “Improving Text-to-Image Alignment with Deep-Fusion Large Language Models”, i cili dokument mund t\u00eb shkarkohet k\u00ebtu: [ https:\/\/arxiv.org\/html\/2409.10695v1<\/a> ] dhe dmth. nga Shtatori i vitit t\u00eb kaluar s’i ka as 1 vit, m\u00eb kujtohet si SOT se \u00e7far\u00eb ishte kualiteti i FOTOREALIZMIT vitin e kaluar, dhe dmth. krahasuar me \u00e7far\u00eb \u00ebsht\u00eb tani, them se v\u00ebrtet p\u00ebr t’u mahnitur sa shpejt zhvillohen dhe avancohen disa gj\u00ebra. <\/p>\n

Me ndihm\u00ebn e foto-realizmit t\u00eb portreteve thuase mund t’i dokumentoj ndryshimet, pasi jan\u00eb ndryshime q\u00eb pothuaj sikur i kam “prekur” vet\u00eb. E pata lexuar pjes\u00ebrisht publikimin e tyre, jo se m\u00eb duhej gj\u00eb por m\u00eb shum\u00eb nga kureshtia lidhur me zhvillimet marramend\u00ebse t\u00eb inteligjenc\u00ebs artifi\u00e7a. At\u00ebbot\u00eb qe b\u00ebr\u00eb rreth gjysm\u00eb viti, apo 1 vit q\u00eb pata filluar me eksperimentet (aty-k\u00ebtu, nga gjysma e 2023), e sakt\u00eb se ekzistonin dhe m\u00eb her\u00ebt (gjat\u00eb 2022) por AI-gjenerimi i imazheve ishte si\u00e7 ia themi n\u00eb “pelenat” e zhvillimit, ende nuk ishin avancuar p\u00ebr t\u00eb qen\u00eb ndonj\u00eb veg\u00ebl interesante.<\/p>\n

Tashm\u00eb “Playground” nuk \u00ebsht\u00eb aktive n\u00eb kuptimin si\u00e7 ishte deri at\u00ebher\u00eb, kompania sikur e pati ndryshuar “kursin”, tashm\u00eb nuk mund t\u00eb gjenerohen imazhet direkt, k\u00ebrkimet dhe zhvillimin e AI p\u00ebr imazhet kompania e ka orientuar kah nj\u00eb model ndryshe i biznesit. Portretet ngjitur k\u00ebtu p\u00ebr mua jan\u00eb “d\u00ebshmia” e k\u00ebsaj mahnitjes q\u00eb e ceka rreth shpejt\u00ebsis\u00eb – prov\u00eb nga un\u00eb vet\u00eb duke e p\u00ebrdorur fuqin\u00eb e IA-s, them “prov\u00eb” q\u00ebkur pothuaj sikur e di kur ishte n\u00eb “pelena”, pastaj n\u00eb “brekushe” dhe shih pastaj… n\u00ebse t’na ishte IA ndonj\u00eb person i men\u00e7ur, ndonj\u00eb shkenc\u00ebtar, vitin e kaluar do i takonte Nobel i mir\u00ebfillt\u00eb : ) <\/p>\n

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

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

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

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

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

 
\n[ fjala e fundit e IA p\u00ebr foto-realizmin<\/strong> ]
\n…her\u00ebn e kaluar pata b\u00ebr\u00eb pak dhe shaka, k\u00ebsaj radhe jo : )
\n(sg – sky division, 3 prill 2025)<\/em><\/p>\n

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

[ fjala e fundit e IA p\u00ebr foto-realizmin<\/strong> ]
\n(sg – sky division, 23 mars 2025)<\/em><\/p>\n

\"\"<\/a>
\n 
\n <\/p>\n","protected":false},"excerpt":{"rendered":"

(sg – sky division, 9 prill 2025 – ca portrete, dhe koment) (…n\u00ebse t’na ishte IA ndonj\u00eb person i men\u00e7ur, ndonj\u00eb shkenc\u00ebtar, vitin e kaluar do i takonte Nobel i mir\u00ebfillt\u00eb : ) Them se disa segmente t\u00eb shkenc\u00ebs e… Lexo →<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[24],"tags":[],"class_list":["post-17072","post","type-post","status-publish","format-standard","hentry","category-skyd"],"_links":{"self":[{"href":"https:\/\/letrat.eu\/index.php?rest_route=\/wp\/v2\/posts\/17072","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/letrat.eu\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/letrat.eu\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/letrat.eu\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/letrat.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=17072"}],"version-history":[{"count":0,"href":"https:\/\/letrat.eu\/index.php?rest_route=\/wp\/v2\/posts\/17072\/revisions"}],"wp:attachment":[{"href":"https:\/\/letrat.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=17072"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/letrat.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=17072"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/letrat.eu\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=17072"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}