{"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\/az\/blog\/news\/grafik-gartnera-2019-o-chyom-vse-eti-modnye-slova","title":{"rendered":"Gartner Qrafiki 2019: bu d\u0259bd\u0259b\u0259li s\u00f6zl\u0259r n\u0259dir?","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

Gartner qrafiki texnologiya sah\u0259sind\u0259 \u00e7al\u0131\u015fanlar \u00fc\u00e7\u00fcn y\u00fcks\u0259k d\u0259b s\u0259rgisin\u0259 b\u0259nz\u0259yir. Ona baxd\u0131\u011f\u0131n\u0131z zaman, bu m\u00f6vs\u00fcm\u00fcn \u0259n trend s\u00f6zl\u0259rini \u00f6nc\u0259d\u0259n \u00f6yr\u0259n\u0259 v\u0259 g\u0259l\u0259c\u0259k konfranslarda n\u0259l\u0259ri e\u015fid\u0259c\u0259yinizi bil\u0259 bil\u0259rsiniz.<\/p>\n

Biz bu qrafikd\u0259ki g\u00f6z\u0259l s\u00f6zl\u0259rin arxas\u0131nda n\u0259l\u0259rin dayand\u0131\u011f\u0131n\u0131 a\u00e7\u0131qlad\u0131q ki, siz d\u0259 bu dild\u0259 dan\u0131\u015fa bil\u0259siniz.<\/p>\n

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

Ba\u015flamaq \u00fc\u00e7\u00fcn bu qrafikin n\u0259 oldu\u011funu siz\u0259 k\u0259skin \u015f\u0259kild\u0259 izah ed\u0259k. H\u0259r il avqust ay\u0131nda Gartner konsaltinq \u015firk\u0259ti - Gartner Hype Curve adl\u0131 hesabat d\u0259rc edir. Bu, az\u0259rbaycanca \"h\u0259y\u0259can qrafiki\" dem\u0259kdir, y\u0259ni sad\u0259 dill\u0259 - h\u0259y\u0259can. 30 il \u0259vv\u0259l Public Enemy qrupu deyirdi: \"H\u0259y\u0259canlara inanmay\u0131n\". \u0130nanmaq ya da inanmamaq f\u0259rdi bir m\u0259s\u0259l\u0259dir, amma bu \u0259sas s\u00f6zl\u0259ri bilm\u0259k vacibdir, \u0259g\u0259r siz texnologiya sah\u0259sind\u0259 i\u015fl\u0259yirsinizs\u0259 v\u0259 qlobal tendensiyalar\u0131 bilm\u0259k ist\u0259yirsinizs\u0259.<\/p>\n

Bu, m\u00fc\u0259yy\u0259n bir texnologiyaya olan c\u0259miyy\u0259tin g\u00f6zl\u0259ntil\u0259rinin qrafikidir. Gartner\u0259 g\u00f6r\u0259, ideal halda, texnologiya ard\u0131c\u0131l olaraq 5 m\u0259rh\u0259l\u0259d\u0259n ke\u00e7ir: texnologiyan\u0131n launching m\u0259rh\u0259l\u0259si, artan g\u00f6zl\u0259ntil\u0259r m\u0259rh\u0259l\u0259si, m\u0259yusluq vadisi, ayd\u0131nla\u015fma meyli v\u0259 m\u0259hsuldarl\u0131q d\u00fczeyi. Lakin b\u0259z\u0259n o, \"m\u0259yusluq vadisi\"nd\u0259 itib bil\u0259r - m\u0259s\u0259l\u0259n, bitcoinl\u0259ri d\u00fc\u015f\u00fcn\u00fcn: \u0259vv\u0259ll\u0259r \"g\u0259l\u0259c\u0259yin pulu\" kimi zirv\u0259y\u0259 \u00e7\u0131xan bu m\u00fcv\u0259ff\u0259qiyy\u0259t, texnologiyan\u0131n \u00e7at\u0131\u015fmazl\u0131qlar\u0131 a\u015fkar olduqca, x\u00fcsus\u0259n d\u0259 transaksiyalar\u0131n say\u0131 il\u0259 ba\u011fl\u0131 m\u0259hdudiyy\u0259tl\u0259r v\u0259 bitcoinl\u0259rin istehsal\u0131 \u00fc\u00e7\u00fcn laz\u0131m olan y\u00fcks\u0259k enerji t\u0259chizat\u0131 (bu da ekoloji probleml\u0259r yarad\u0131r) il\u0259 a\u015fa\u011f\u0131 d\u00fc\u015fd\u00fc. V\u0259 \u0259lb\u0259tt\u0259 ki, unutmaq olmaz ki, Gartner qrafiki sad\u0259c\u0259 bir proqnozdur: burada, m\u0259s\u0259l\u0259n, \u0259n parlaq realla\u015fmam\u0131\u015f proqnozlar\u0131 ara\u015fd\u0131ran detallar\u0131 oxuya bil\u0259rsiniz. m\u0259qal\u0259<\/a><\/noindex>, harada \u0259n parlaq reall\u0131qda ba\u015f verm\u0259y\u0259n proqnozlar bar\u0259d\u0259 dan\u0131\u015f\u0131l\u0131r.<\/p>\n

Bel\u0259likl\u0259, Gartner qrafikini n\u0259z\u0259rd\u0259n ke\u00e7ir\u0259k. Texnologiyalar 5 \u0259sas m\u00f6vzu qrupuna b\u00f6l\u00fcnm\u00fc\u015fd\u00fcr:<\/p>\n

    \n
  1. \u0130r\u0259lid\u0259ki AI v\u0259 analitika (Advanced AI and Analytics)\n<\/li>\n
  2. Postklasik hesablama v\u0259 kommunikasiyalar (Postclassical Compute and Comms)\n<\/li>\n
  3. Sensorlar v\u0259 mobilitet (Sensing and Mobility)\n<\/li>\n
  4. \"Tamamlanm\u0131\u015f\" insan (Augmented Human)\n<\/li>\n
  5. R\u0259q\u0259msal ekosisteml\u0259r (Digital Ecosystems)\n<\/li>\n<\/ol>\n

    <\/p>\n

    1. \u0130r\u0259lid\u0259ki AI v\u0259 analitika (Advanced AI and Analytics)<\/h1>\n

    \nSon 10 ild\u0259 d\u0259rin \u00f6yr\u0259nm\u0259nin ulduz d\u00f6vr\u00fcn\u00fc g\u00f6r\u00fcr\u00fck. Bu \u015f\u0259b\u0259k\u0259l\u0259r h\u0259qiq\u0259t\u0259n \u00f6z sah\u0259l\u0259rind\u0259 effektivdir. 2018-ci ild\u0259 Yan LeCun, Ceffri Hintton v\u0259 Yosua Benjiodan d\u0259rin \u00f6yr\u0259nm\u0259y\u0259 dair a\u015fkarlar\u0131na g\u00f6r\u0259 Turing m\u00fckafat\u0131 \u2013 informatika sah\u0259sind\u0259ki Nizamnam\u0259y\u0259 b\u0259nz\u0259r \u0259n prestijli m\u00fckafat\u0131 ald\u0131lar. Bel\u0259likl\u0259, bu sah\u0259d\u0259ki \u0259sas tendensiyalar qrafikd\u0259 t\u0259qdim olunmu\u015fdur:<\/p>\n

    1.1. \u00d6yr\u0259nm\u0259 transferi (Transfer Learning)<\/h2>\n

    \nSiz neyron \u015f\u0259b\u0259k\u0259sini s\u0131f\u0131rdan \u00f6yr\u0259tmirsiniz, art\u0131q \u00f6yr\u0259dilmi\u015f birini g\u00f6t\u00fcr\u0259r\u0259k ona yeni bir m\u0259qs\u0259d t\u0259yin edirsiniz. B\u0259z\u0259n bunun \u00fc\u00e7\u00fcn \u015f\u0259b\u0259k\u0259nin bir hiss\u0259sini yenid\u0259n \u00f6yr\u0259tm\u0259k laz\u0131m g\u0259l\u0259 bil\u0259r, amma b\u00fct\u00fcn \u015f\u0259b\u0259k\u0259ni deyil, bu is\u0259 daha s\u00fcr\u0259tlidir. M\u0259s\u0259l\u0259n, haz\u0131r ResNet50 neyron \u015f\u0259b\u0259k\u0259sini ImageNet1000 veril\u0259nl\u0259r d\u0259stind\u0259 \u00f6yr\u0259d\u0259r\u0259k g\u00f6t\u00fcrs\u0259niz, \u015f\u0259kil \u0259sas\u0131nda \u00e7ox sayda m\u00fcxt\u0259lif obyektl\u0259ri d\u0259rin bir s\u0259viyy\u0259d\u0259 (50 qatl\u0131 neyron \u015f\u0259b\u0259k\u0259si t\u0259r\u0259find\u0259n haz\u0131rlanm\u0131\u015f meyarlara g\u00f6r\u0259 1000 sinif) t\u0259snif ed\u0259 bil\u0259n bir alqoritm \u0259ld\u0259 ed\u0259rsiniz. Ancaq b\u00fct\u00fcn bu \u015f\u0259b\u0259k\u0259ni tamamil\u0259 \u00f6yr\u0259tm\u0259y\u0259 ehtiyac yoxdur ki, bu da aylar ala bil\u0259r.<\/p>\n

    \u0130 onlayn-kurs<\/a><\/noindex> Samsung \u00abNeyron \u015f\u0259b\u0259k\u0259l\u0259ri v\u0259 komp\u00fcter g\u00f6rm\u0259si\u00bb, n\u00fcmun\u0259 \u00fc\u00e7\u00fcn, finalda Kaggle-tap\u015f\u0131r\u0131\u011f\u0131<\/a><\/noindex> t\u0259miz v\u0259 \u00e7irkli bo\u015fqablar\u0131n t\u0259snifat\u0131 il\u0259, 5 d\u0259qiq\u0259 \u0259rzind\u0259 siz\u0259 d\u0259rin neyron \u015f\u0259b\u0259k\u0259 t\u0259qdim ed\u0259n bir yana\u015fma n\u00fcmayi\u015f etdirilir, bu \u015f\u0259b\u0259k\u0259 yuxar\u0131da t\u0259svir olunmu\u015f arxitektura \u0259sas\u0131nda \u00e7irkli bo\u015fqablar\u0131 t\u0259mizl\u0259rd\u0259n ay\u0131rma\u011f\u0131 \u00f6yr\u0259nir. \u0130btidai \u015f\u0259b\u0259k\u0259 he\u00e7 vaxt bo\u015fqablar\u0131n n\u0259 oldu\u011funu bilmirdi, yaln\u0131z qu\u015flar\u0131 itl\u0259rd\u0259n ay\u0131rma\u011f\u0131 \u00f6yr\u0259nmi\u015fdi (bax\u0131n ImageNet).<\/p>\n

    \"Gartner
    \nM\u0259nb\u0259:     onlayn-kurs<\/a><\/noindex> Samsung \u00abNeyron \u015f\u0259b\u0259k\u0259l\u0259ri v\u0259 komp\u00fcter g\u00f6rm\u0259si\u00bb<\/i><\/p>\n

    Transfer Learning \u00fc\u00e7\u00fcn hans\u0131 yana\u015fmalar\u0131n i\u015fl\u0259diyini v\u0259 hans\u0131 m\u00f6vcud \u0259sas arxitekturalar\u0131n oldu\u011funu bilm\u0259k laz\u0131md\u0131r. \u00dcmumiyy\u0259tl\u0259, bu, ma\u015f\u0131n \u00f6yr\u0259nm\u0259nin praktiki t\u0259tbiql\u0259rinin yaranmas\u0131n\u0131 \u00e7ox s\u00fcr\u0259tl\u0259ndirir.<\/p>\n

    1.2. Generativ-t\u0259zyiq \u015f\u0259b\u0259k\u0259l\u0259ri (Generative Adversarial Networks, GAN)<\/h2>\n

    \nBu, biz\u0259 \u00f6yr\u0259nm\u0259 m\u0259qs\u0259dini formala\u015fd\u0131rman\u0131n \u00e7ox \u00e7\u0259tin oldu\u011fu hallard\u0131r. M\u0259s\u0259l\u0259 realla\u015fan h\u0259yata n\u0259 q\u0259d\u0259r yax\u0131n olsa, o q\u0259d\u0259r daha ayd\u0131nd\u0131r (\u00ab\u015fkaf\u0131 g\u0259tir\u00bb), amma onu texniki tap\u015f\u0131r\u0131q kimi formala\u015fd\u0131rmaq bir o q\u0259d\u0259r \u00e7\u0259tindir. GAN - bizi bu problemi h\u0259ll etm\u0259k istiqam\u0259tind\u0259 d\u0131rnaqqan bir c\u0259hd.<\/p>\n

    Burada iki \u015f\u0259b\u0259k\u0259 i\u015fl\u0259yir: biri generator (Generative), dig\u0259ri is\u0259 diskriminator (Adversarial). Bir \u015f\u0259b\u0259k\u0259 faydal\u0131 i\u015f g\u00f6rm\u0259yi \u00f6yr\u0259nir (\u015f\u0259kill\u0259ri t\u0259snif etm\u0259k, s\u0259sl\u0259ri tan\u0131maq, cizgi filml\u0259ri \u00e7\u0259km\u0259k). Dig\u0259r \u015f\u0259b\u0259k\u0259 is\u0259 o \u015f\u0259b\u0259k\u0259ni \u00f6yr\u0259tm\u0259yi \u00f6yr\u0259nir: onun real n\u00fcmun\u0259l\u0259ri var v\u0259 o, generator hiss\u0259sinin ortaya qoydu\u011fu m\u0259hsullar\u0131 real d\u00fcnya obyektl\u0259ri (\u00f6yr\u0259tm\u0259 d\u0259stl\u0259ri) il\u0259 m\u00fcqayis\u0259 etm\u0259k \u00fc\u00e7\u00fcn \u0259vv\u0259lc\u0259d\u0259n bilinm\u0259y\u0259n m\u00fcr\u0259kk\u0259b bir formulu tapma\u011f\u0131 \u00f6yr\u0259nir, \u0259h\u0259miyy\u0259tli d\u0259rin x\u00fcsusiyy\u0259tl\u0259rl\u0259: g\u00f6zl\u0259rin say\u0131, Miyazaki stilin\u0259 yax\u0131nl\u0131q, ingilis dili t\u0259l\u0259ff\u00fcz\u00fcn\u00fcn d\u00fczg\u00fcnl\u00fcy\u00fc.<\/p>\n

    \"Gartner
    \n\u015e\u0259b\u0259k\u0259nin anime xarakterl\u0259ri yaratma n\u0259tic\u0259sinin n\u00fcmun\u0259si.     M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    Amma orada, \u0259lb\u0259tt\u0259 ki, arxitekturan\u0131 qura\u015fd\u0131rmaq \u00e7\u0259tindir. Sad\u0259c\u0259 neyronlar\u0131 atmaq kifay\u0259t deyil, onlar\u0131 haz\u0131rlamaq laz\u0131md\u0131r. V\u0259 \u00f6yr\u0259tm\u0259k bir ne\u00e7\u0259 h\u0259ft\u0259 \u00e7\u0259kir. GAN il\u0259 ba\u011fl\u0131 m\u0259s\u0259l\u0259l\u0259rl\u0259 Samsung-in S\u00fcni \u0130ntellekt M\u0259rk\u0259zind\u0259ki h\u0259mkarlar\u0131m m\u0259\u015f\u011fuldur, bu onlara \u0259sas t\u0259dqiqat suallar\u0131ndan biridir. M\u0259s\u0259l\u0259n, bel\u0259 bir inki\u015faf<\/a><\/noindex>: istifad\u0259\u00e7il\u0259rin d\u0259yi\u015fdiril\u0259 bil\u0259n pozalarda insanlar\u0131n reall\u0131\u011fa yax\u0131n fotolar\u0131n\u0131 sintez etm\u0259k \u00fc\u00e7\u00fcn generativ \u015f\u0259b\u0259k\u0259l\u0259rin istifad\u0259si \u2014 m\u0259s\u0259l\u0259n, virtual geyim kabin\u0259si yaratmaq \u00fc\u00e7\u00fcn, ya da \u00fcz sintezi \u00fc\u00e7\u00fcn ki, bu, keyfiyy\u0259tli videokonfranslar, yay\u0131m yaxud \u015f\u0259xsi m\u0259lumatlar\u0131 qorumaq \u00fc\u00e7\u00fcn saxlanmal\u0131 ya da \u00f6t\u00fcr\u00fclm\u0259li olan m\u0259lumatlar\u0131n miqdar\u0131n\u0131 azaltma\u011fa imkan ver\u0259 bil\u0259r.<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    1.3. \u0130zah edil\u0259n AI (Explainable AI)<\/h2>\n

    \nB\u0259zi nadir m\u0259s\u0259l\u0259l\u0259rd\u0259 d\u0259rin arxitekturalarda ir\u0259lil\u0259yi\u015f, d\u0259rin s\u00fcni neyron \u015f\u0259b\u0259k\u0259l\u0259rinin imkanlar\u0131n\u0131 insanlara yax\u0131nla\u015fd\u0131rd\u0131. \u0130ndi m\u00fcbariz\u0259 bel\u0259 m\u0259s\u0259l\u0259l\u0259rin dair\u0259sini geni\u015fl\u0259ndirm\u0259k \u00fc\u00e7\u00fcnd\u00fcr. M\u0259s\u0259l\u0259n, bir robot tozsoran, qar\u015f\u0131la\u015fd\u0131qda asanl\u0131qla pi\u015fiyi itd\u0259n ay\u0131ra bil\u0259r. Amma \u0259ks\u0259r h\u0259yati situasiyalarda o, paltar v\u0259 ya mebel aras\u0131nda yatan pi\u015fiyi tapa bilmir (t\u0259bii ki, biz d\u0259 \u0259ks\u0259r hallarda bel\u0259…).<\/p>\n

    D\u0259rin neyron \u015f\u0259b\u0259k\u0259l\u0259rinin u\u011furlar\u0131n\u0131n s\u0259b\u0259bi n\u0259dir? Onlar tap\u015f\u0131r\u0131\u011f\u0131n t\u0259svirini \u2018g\u00f6zl\u0259 g\u00f6r\u00fcn\u0259n\u2019 m\u0259lumatlara (foto\u015f\u0259kil piksel\u0259ri, s\u0259sin dal\u011falanmalar\u0131…) \u0259saslanaraq formala\u015fd\u0131rm\u0131r, \u0259ksin\u0259, bu m\u0259lumatlar\u0131n bir ne\u00e7\u0259 y\u00fcz neyron \u015f\u0259b\u0259k\u0259si qat\u0131 il\u0259 \u00f6nc\u0259ki emal\u0131ndan sonra al\u0131nm\u0131\u015f \u00f6z\u0259llikl\u0259r\u0259 \u0259saslan\u0131r. T\u0259\u0259ss\u00fcf ki, bu \u0259laq\u0259l\u0259r h\u0259m d\u0259 m\u0259nas\u0131z, ziddiyy\u0259tli ola bil\u0259r v\u0259 ilkin veril\u0259nl\u0259r bazas\u0131n\u0131n qeyri-m\u00fck\u0259mm\u0259lliyinin izl\u0259rini da\u015f\u0131ya bil\u0259r. M\u0259s\u0259l\u0259n, AI-nin rekrutinqd\u0259 d\u00fc\u015f\u00fcn\u00fclm\u0259d\u0259n t\u0259tbiqinin n\u0259 il\u0259 n\u0259tic\u0259l\u0259n\u0259 bil\u0259c\u0259yini \u0259ks etdir\u0259n ki\u00e7ik bir komp\u00fcter oyunu var. Survival Of The Best Fit<\/a><\/noindex>.<\/p>\n

    \"Gartner
    \n\u015e\u0259kil etiketl\u0259m\u0259 sistemi, bir \u015f\u0259kil - ger\u00e7\u0259kd\u0259 ki\u015fi olan bir a\u015fpaz\u0131 qad\u0131n kimi adland\u0131rd\u0131 (    M\u0259nb\u0259<\/a><\/noindex>). Bu qeyd edildi<\/a><\/noindex> Virjiniya \u0130nstitutunda.<\/i><\/p>\n

    Bizim formul\u0259 ed\u0259 bilm\u0259diyimiz m\u00fcr\u0259kk\u0259b v\u0259 d\u0259rin \u0259laq\u0259l\u0259ri analiz etm\u0259k \u00fc\u00e7\u00fcn izah edil\u0259n AI metodlar\u0131 laz\u0131md\u0131r. Bu metodlar d\u0259rin neyron \u015f\u0259b\u0259k\u0259l\u0259rinin x\u00fcsusiyy\u0259tl\u0259rini t\u0259\u015fkil edir ki, \u00f6yr\u0259nm\u0259 tamamland\u0131qdan sonra \u015f\u0259b\u0259k\u0259nin \u00f6yr\u0259ndiyi daxili n\u00fcmay\u0259nd\u0259liyi analiz edib, yaln\u0131z onun q\u0259rar\u0131na etibar etm\u0259y\u0259k.<\/p>\n

    1.4. K\u0259nar Analitika \/ AI (Edge Analytics \/ AI)<\/h2>\n

    \nK\u0259nar s\u00f6z\u00fcn\u00fcn oldu\u011fu h\u0259r \u015fey, \u0259slind\u0259, alqoritml\u0259rin bir hiss\u0259sinin bulud\/serverd\u0259n son cihaz\/qap\u0131 s\u0259viyy\u0259sin\u0259 k\u00f6\u00e7\u00fcr\u00fclm\u0259sini ifad\u0259 edir. Bel\u0259 bir alqoritm daha s\u00fcr\u0259tli f\u0259aliyy\u0259t g\u00f6st\u0259r\u0259c\u0259k v\u0259 m\u0259rk\u0259zi server\u0259 qo\u015fulma\u011fa ehtiyac duymayacaq. \u201c\u0130nce m\u00fc\u015ft\u0259ri\u201d m\u0259fhumunu tan\u0131y\u0131rs\u0131n\u0131zsa, burada m\u00fc\u015ft\u0259rini bir az daha ya\u011fl\u0131 edirik.
    \nBu, \u015feyl\u0259r interneti \u00fc\u00e7\u00fcn vacib ola bil\u0259r. M\u0259s\u0259l\u0259n, \u0259g\u0259r ma\u015f\u0131n \u00e7ox isindiyind\u0259 v\u0259 soyutma t\u0259l\u0259b edirs\u0259, d\u0259rhal, zavod s\u0259viyy\u0259sind\u0259 bu bar\u0259d\u0259 siqnal g\u00f6nd\u0259rm\u0259k m\u0259nal\u0131d\u0131r, m\u0259lumat\u0131n buluda d\u00fc\u015fm\u0259sini v\u0259 oradan n\u00f6vb\u0259t\u00e7iy\u0259 \u00e7atmas\u0131n\u0131 g\u00f6zl\u0259m\u0259d\u0259n. Ya da ba\u015fqa bir misal: \u00f6z-\u00f6z\u00fcn\u0259 s\u00fcr\u0259n avtomobill\u0259r, m\u0259rk\u0259zi server\u0259 m\u00fcraci\u0259t etm\u0259d\u0259n, yol \u015f\u0259raitini \u00f6zl\u0259ri h\u0259ll ed\u0259 bil\u0259rl\u0259r.<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    Yada bir t\u0259hl\u00fck\u0259sizlik bax\u0131m\u0131ndan niy\u0259 bunun vacib oldu\u011funu g\u00f6st\u0259r\u0259n ba\u015fqa bir n\u00fcmun\u0259: telefonunuzda m\u0259tnl\u0259ri daxil ed\u0259rk\u0259n, o, sizin \u00fc\u00e7\u00fcn tipik olan s\u00f6zl\u0259ri yadda saxlay\u0131r ki, telefonun klaviaturas\u0131 siz\u0259 onlar\u0131 rahatl\u0131qla g\u00f6st\u0259rsin \u2013 bu, proqnozverici m\u0259tni daxil etm\u0259 adlan\u0131r. Klaviaturada daxil etdiyiniz h\u0259r \u015feyi bir m\u0259lumat m\u0259rk\u0259zin\u0259 g\u00f6nd\u0259rm\u0259k sizin gizliliyinizin pozulmas\u0131 olard\u0131 v\u0259 sad\u0259c\u0259 olaraq t\u0259hl\u00fck\u0259siz deyil. Buna g\u00f6r\u0259 d\u0259, klaviaturan\u0131n \u00f6yr\u0259nilm\u0259si yaln\u0131z sizin cihaz\u0131n\u0131z\u0131n \u00e7\u0259r\u00e7iv\u0259sind\u0259 ba\u015f verir.<\/p>\n

    1.5. AI Platformas\u0131 Xidm\u0259t olaraq (AI PaaS)<\/h2>\n

    \nPaaS \u2013 Xidm\u0259t olaraq Platforma \u2013 bu, inteqrasiya olunmu\u015f bir platformaya, onun bulud m\u0259lumat saxlama v\u0259 haz\u0131r prosedurlar\u0131na daxil olma\u011f\u0131 t\u0259min ed\u0259n biznes modelidir. Buna g\u00f6r\u0259 d\u0259, infrastruktur m\u0259s\u0259l\u0259l\u0259rind\u0259n azad olmaqla, faydal\u0131 bir \u015fey istehsal etm\u0259y\u0259 tamamil\u0259 diqq\u0259timizi y\u00f6n\u0259ld\u0259 bilirik. AI \u00fc\u00e7\u00fcn PaaS platformalar\u0131na n\u00fcmun\u0259l\u0259r: IBM Cloud, Microsoft Azure, Amazon Machine Learning, Google AI Platform.<\/p>\n

    1.6. Adaptiv Ma\u015f\u0131n \u00d6yr\u0259nm\u0259si (Adaptive ML)<\/h2>\n

    \n\u0130kincisi, s\u00fcni intellektin adaptasiya olmas\u0131na icaz\u0259 vers\u0259k... Sizin soru\u015faca\u011f\u0131n\u0131z sual olacaq - nec\u0259 y\u0259ni?.. O, art\u0131q tap\u015f\u0131r\u0131\u011fa uy\u011funla\u015fm\u0131rm\u0131? Problemin m\u0259rk\u0259zind\u0259 budur: bel\u0259 tap\u015f\u0131r\u0131qlar\u0131n h\u0259r birini biz diqq\u0259tl\u0259 haz\u0131r edirik, sonra s\u00fcni intellektin bu tap\u015f\u0131r\u0131\u011f\u0131 yerin\u0259 yetirm\u0259si \u00fc\u00e7\u00fcn bir alqoritm qururuq. \u0130nsanlar siz\u0259 cavab ver\u0259c\u0259k \u2013 m\u0259lum olur ki, bu ard\u0131c\u0131ll\u0131\u011f\u0131 sad\u0259l\u0259\u015fdirm\u0259k d\u0259 m\u00fcmk\u00fcnd\u00fcr.<\/p>\n

    Adi ma\u015f\u0131n \u00f6yr\u0259nm\u0259si a\u00e7\u0131q sistem prinsipi (open-loop) il\u0259 i\u015fl\u0259yir: siz m\u0259lumatlar\u0131 haz\u0131rlay\u0131rs\u0131n\u0131z, neyron \u015f\u0259b\u0259k\u0259sini (ya da ba\u015fqalar\u0131n\u0131) d\u00fc\u015f\u00fcn\u00fcrs\u00fcn\u00fcz, \u00f6yr\u0259dirsiniz, sonra bir ne\u00e7\u0259 g\u00f6st\u0259riciy\u0259 bax\u0131rs\u0131n\u0131z, v\u0259 \u0259g\u0259r h\u0259r \u015fey b\u0259y\u0259ndiyiniz \u015f\u0259kild\u0259dirs\u0259, neyron \u015f\u0259b\u0259k\u0259sini smartfonlara g\u00f6nd\u0259rm\u0259k olar \u2013 istifad\u0259\u00e7il\u0259rin tap\u015f\u0131r\u0131qlar\u0131n\u0131 h\u0259ll etm\u0259k \u00fc\u00e7\u00fcn. Lakin, m\u0259lumatlar\u0131n \u00e7ox oldu\u011fu v\u0259 onlar\u0131n xarakterinin t\u0259dric\u0259n d\u0259yi\u015fdiyi hallarda, dig\u0259r metodlar t\u0259l\u0259b olunur. Bu c\u00fcr sisteml\u0259r, \u00f6zl\u0259rini adaptasiya ed\u0259n v\u0259 \u00f6zl\u0259rini \u00f6yr\u0259d\u0259n, ba\u011flanm\u0131\u015f, \u00f6z-\u00f6z\u00fcn\u0259 \u00f6yr\u0259n\u0259n d\u00f6vrl\u0259rd\u0259 t\u0259\u015fkil olunur (closed-loop) v\u0259 davaml\u0131 i\u015fl\u0259m\u0259lidir.<\/p>\n

    T\u0259tbiql\u0259r - bu, \u00e7ox sayda biznesmenl\u0259rin q\u0259rar verdiyi ax\u0131n analitikas\u0131 (Stream Analytics) ola bil\u0259r, ya da adaptiv istehsal idar\u0259\u00e7iliyi. M\u00fcasir t\u0259tbiql\u0259rin miqyas\u0131nda v\u0259 insanlar\u0131n daha yax\u015f\u0131 ba\u015fa d\u00fc\u015fd\u00fcy\u00fc riskl\u0259r n\u0259z\u0259r\u0259 al\u0131nd\u0131qda, bu problemin h\u0259llini t\u0259qdim ed\u0259n metodlar, b\u00fct\u00fcn bu metodlar Adaptiv AI \u00fcmumi ad\u0131 alt\u0131nda toplan\u0131r.<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    Bu \u015f\u0259kil \u00fcz\u0259rind\u0259 baxanda, futuristl\u0259rin \u00e7\u00f6r\u0259ksiz qalmad\u0131\u011f\u0131n\u0131 d\u00fc\u015f\u00fcnm\u0259k \u00e7\u0259tindir \u2013 robotun n\u0259f\u0259s alma\u011f\u0131na \u00f6yr\u0259d\u0259nd\u0259...<\/p>\n

    Postklasik hesablama v\u0259 kommunikasiyalar (Postclassical Compute and Comms)<\/h1>\n

    <\/p>\n

    2.1. Be\u015finci N\u0259sil Mobil Rabit\u0259 (5G)<\/h2>\n

    \nBu q\u0259d\u0259r maraql\u0131 bir m\u00f6vzudur ki, d\u0259rhal siz\u0259 bizim makalede<\/a><\/noindex>. Burada is\u0259 q\u0131sa bir x\u00fclas\u0259 var. 5G, m\u0259lumatlar\u0131n \u00f6t\u00fcr\u00fclm\u0259 tezliyinin art\u0131r\u0131lmas\u0131 il\u0259 \u0130nternetin s\u00fcr\u0259tini qeyri-realist s\u00fcr\u0259tli ed\u0259c\u0259k. Q\u0131sa dal\u011falar\u0131n mane\u0259l\u0259rd\u0259n ke\u00e7m\u0259si \u00e7\u0259tindir, buna g\u00f6r\u0259 d\u0259 \u015f\u0259b\u0259k\u0259l\u0259rin qurulmas\u0131 tamamil\u0259 f\u0259rqli olacaq: baza stansiyalar\u0131n\u0131n say\u0131 500 d\u0259f\u0259 daha \u00e7ox olmal\u0131d\u0131r.<\/p>\n

    S\u00fcr\u0259tl\u0259 birlikd\u0259 bizim yeni fenomenl\u0259rimiz olacaq: reall\u0131qda \u0259yl\u0259nc\u0259 oyunlar\u0131, teleprezensiya vasit\u0259sil\u0259 m\u00fcr\u0259kk\u0259b tap\u015f\u0131r\u0131qlar\u0131n (c\u0259rrahiyy\u0259 kimi) icras\u0131, avtomobill\u0259r aras\u0131nda \u00fcnsiyy\u0259t vasit\u0259sil\u0259 q\u0259zalar\u0131n v\u0259 \u00e7\u0259tin v\u0259ziyy\u0259tl\u0259rin qar\u015f\u0131s\u0131n\u0131n al\u0131nmas\u0131. Daha adi bir m\u0259qamda: k\u00fctl\u0259vi t\u0259dbirl\u0259r zaman\u0131, m\u0259s\u0259l\u0259n, stadionlardak\u0131 mat\u00e7larda mobil internetin d\u00fc\u015fm\u0259si n\u0259hay\u0259t dayanacaq.<\/p>\n

    \"Gartner
    \n\u015e\u0259kil m\u0259nb\u0259yi \u2014 Reuters, Niantic<\/i><\/p>\n

    2.2. N\u00f6vb\u0259ti N\u0259sil Yadda\u015f (Next-Generation Memory)<\/h2>\n

    \nBurada be\u015finci n\u0259sil operativ yadda\u015f \u2013 DDR5 haqq\u0131nda dan\u0131\u015f\u0131l\u0131r. Samsung 2019-cu ilin sonuna q\u0259d\u0259r DDR5 \u0259sasl\u0131 m\u0259hsullar\u0131n t\u0259qdim ed\u0259c\u0259yini a\u00e7\u0131qlad\u0131. Yeni yadda\u015f\u0131n iki d\u0259f\u0259 daha s\u00fcr\u0259tli v\u0259 iki d\u0259f\u0259 daha \u00e7ox y\u00fckl\u0259m\u0259 tutumuna sahib olaca\u011f\u0131 g\u00f6zl\u0259nilir, y\u0259ni komp\u00fcterimiz \u00fc\u00e7\u00fcn 32GB-a q\u0259d\u0259r yadda\u015f modullar\u0131 \u0259ld\u0259 ed\u0259 bil\u0259c\u0259yik. G\u0259l\u0259c\u0259kd\u0259 bu, smartfonlar \u00fc\u00e7\u00fcn x\u00fcsusil\u0259 aktualla\u015facaq (yeni yadda\u015f a\u015fa\u011f\u0131 enerji istehlak\u0131na malik versiyada olacaq) v\u0259 noutbuklar \u00fc\u00e7\u00fcn (DIMM slotlar\u0131n\u0131n say\u0131 m\u0259hduddur). Eyni zamanda, ma\u015f\u0131n \u00f6yr\u0259nm\u0259si b\u00f6y\u00fck m\u0259bl\u0259\u011fd\u0259 operativ yadda\u015f t\u0259l\u0259b edir.<\/p>\n

    2.3. A\u015fa\u011f\u0131-Elektrik Orbitli Peyk Sisteml\u0259ri (Low-Earth-Orbit Satellite Systems)<\/h2>\n

    \nB\u00f6y\u00fck, baha v\u0259 g\u00fccl\u00fc peykl\u0259rin ki\u00e7ik v\u0259 ucuz kolonlarla \u0259v\u0259z edilm\u0259 ideyas\u0131 90-c\u0131 ill\u0259rd\u0259n b\u0259ri tan\u0131\u015fd\u0131r. H\u0259tta... \u00abElon Musk tezlikl\u0259 h\u0259r k\u0259s\u0259 peyk vasit\u0259sil\u0259 internet ver\u0259c\u0259k\u00bb<\/a><\/noindex> indi bunu e\u015fitm\u0259y\u0259n yaln\u0131z t\u0259nb\u0259ll\u0259rdir. Bu sah\u0259d\u0259ki \u0259n tan\u0131nm\u0131\u015f \u015firk\u0259t Iridium'dur, 90-c\u0131 ill\u0259rin sonunda iflas etdi, lakin AB\u015e M\u00fcdafi\u0259 Nazirliyi t\u0259r\u0259find\u0259n xilas edildi (iRidium \u2013 Rusiya a\u011f\u0131ll\u0131 ev sistemi il\u0259 qar\u0131\u015fd\u0131rmay\u0131n). Elon Musk-\u0131n layih\u0259si (Starlink) t\u0259k deyil \u2013 peyk yar\u0131\u015f\u0131nda Richard Branson (OneWeb \u2013 1440 potensial peyk), Boeing (3000 peyk), Samsung (4600 peyk) v\u0259 ba\u015fqalar\u0131 i\u015ftirak edirl\u0259r.<\/p>\n

    Bu sah\u0259d\u0259 v\u0259ziyy\u0259t nec\u0259dir, iqtisadiyyat\u0131 n\u0259 c\u00fcrd\u00fcr - bunu oxuyun bax\u0131\u015fda<\/a><\/noindex>. Biz is\u0259 bu sisteml\u0259rin ilk istifad\u0259\u00e7il\u0259ri t\u0259r\u0259find\u0259n h\u0259yata ke\u00e7iril\u0259c\u0259k ilk testl\u0259rini g\u00f6zl\u0259yirik ki, bunlar art\u0131q g\u0259l\u0259n il ba\u015f tutmal\u0131d\u0131r.<\/p>\n

    2.4. Nanomassada 3D \u00c7ap (Nanoscale 3D Printing)<\/h2>\n

    \n3D \u00e7ap, h\u0259r bir insan\u0131n h\u0259yat\u0131na daxil olmasa da (\u015f\u0259xsi ev plastik fabriki formas\u0131nda v\u0259d edils\u0259 d\u0259), art\u0131q texnologiya ni\u015find\u0259n \u00e7\u0131xm\u0131\u015fd\u0131r. Bunun g\u00f6st\u0259ricisi, \u0259n az\u0131ndan 3D heyk\u0259l \u00e7\u0259k\u0259n q\u0259l\u0259ml\u0259rin m\u00f6vcud oldu\u011funun h\u0259r bir m\u0259kt\u0259blinin bildiyi, bir \u00e7oxunun \u00f6zl\u0259rin\u0259 \u00e7\u0259kili\u015f v\u0259 ekstruder qutusu almaq arzusu oldu\u011fu ger\u00e7\u0259kl\u0259ridir... \u201csad\u0259c\u0259 el\u0259\u201d (ya da art\u0131q alm\u0131\u015fd\u0131rlar).<\/p>\n

    Stereolitografiya (laserli 3D printerl\u0259r) ayr\u0131-ayr\u0131 fotonlarla \u00e7ap etm\u0259y\u0259 imkan verir: iki fotonun s\u0259rtl\u0259\u015fm\u0259si \u00fc\u00e7\u00fcn kifay\u0259t q\u0259d\u0259r yeni polimerl\u0259r ara\u015fd\u0131r\u0131l\u0131r. Bu, laboratoriya \u015f\u0259raitind\u0259n k\u0259narda tamamil\u0259 yeni filtr, tutucular, yaylar, kapilyar, linzalar v\u0259... \u015f\u0259rhl\u0259rd\u0259 sizin variantlar yaradacaqd\u0131r! V\u0259 burada fotopolymerizasiya prosesin\u0259 yax\u0131nla\u015f\u0131r\u0131q - yaln\u0131z bu texnologiya \u00ab\u00e7ap etm\u0259y\u0259\u00bb imkan verir prosessorlar v\u0259 hesablama sxeml\u0259ri. Bundan \u0259lav\u0259, art\u0131q ill\u0259rdir var 500-nm qrafen \u00fc\u00e7 \u00f6l\u00e7\u00fcl\u00fc strukturlar\u0131n \u00e7ap texnologiyas\u0131<\/a><\/noindex>, lakin radikal inki\u015faf olmadan.<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259<\/i><\/a><\/noindex><\/p>\n

    3. Sensorika v\u0259 h\u0259r\u0259k\u0259tlilik (Sensing and Mobility)<\/h1>\n

    <\/p>\n

    3.1. Avtonom avtomobill\u0259r, s\u0259viyy\u0259 4 v\u0259 5 (Autonomous Driving Level 4 & 5)<\/h2>\n

    \nTerminologiyada qar\u0131\u015fmamaq \u00fc\u00e7\u00fcn, avtonomluq s\u0259viyy\u0259l\u0259rini ba\u015fa d\u00fc\u015fm\u0259k laz\u0131md\u0131r (daha \u0259trafl\u0131 \u00fc\u00e7\u00fcn makaleler<\/a><\/noindex>, maraqlananlara g\u00f6nd\u0259rdiyimiz):<\/p>\n

    S\u0259viyy\u0259 1: Kruiz-kontrol: s\u00fcr\u00fcc\u00fcy\u0259 \u00e7ox m\u0259hdud v\u0259ziyy\u0259tl\u0259rd\u0259 k\u00f6m\u0259k (m\u0259s\u0259l\u0259n, s\u00fcr\u00fcc\u00fc pedal\u0131n \u00fcst\u00fcnd\u0259n aya\u011f\u0131n\u0131 g\u00f6t\u00fcrd\u00fckd\u0259n sonra avtomobili m\u00fc\u0259yy\u0259n s\u00fcr\u0259td\u0259 saxlama\u011fa k\u00f6m\u0259k edilm\u0259si)
    \nS\u0259viyy\u0259 2: Y\u00f6nl\u0259ndirm\u0259 v\u0259 tormozlamada m\u0259hdud k\u00f6m\u0259k. S\u00fcr\u00fcc\u00fc dem\u0259k olar ki, d\u0259rhal idar\u0259ni g\u00f6t\u00fcrm\u0259y\u0259 haz\u0131r olmal\u0131d\u0131r. Onun \u0259li s\u00fckan\u0131n \u00fcst\u00fcnd\u0259, bax\u0131\u015f\u0131 yolda. Bu art\u0131q Tesla v\u0259 General Motors-da var.
    \nS\u0259viyy\u0259 3: S\u00fcr\u00fcc\u00fcn\u00fcn art\u0131q yola daim baxmas\u0131 laz\u0131m deyil. Ancaq o, diqq\u0259tli qalmal\u0131 v\u0259 idar\u0259ni g\u00f6t\u00fcrm\u0259y\u0259 haz\u0131r olmal\u0131d\u0131r. Bu, haz\u0131rda sat\u0131\u015fda olan avtomobill\u0259rd\u0259 yoxdur. Haz\u0131rda m\u00f6vcud olanlar - s\u0259viyy\u0259 1-2 aras\u0131ndad\u0131r.
    \nS\u0259viyy\u0259 4: H\u0259qiqi avtopilot, lakin m\u0259hdudiyy\u0259tl\u0259rl\u0259: yaln\u0131z m\u00fcvafiq sah\u0259l\u0259rd\u0259, tam kartoqrafiya edilmi\u015f v\u0259 sistem\u0259 m\u0259lum olan sah\u0259l\u0259rd\u0259 v\u0259 m\u00fc\u0259yy\u0259n \u015f\u0259rtl\u0259rd\u0259: m\u0259s\u0259l\u0259n, qar olmamas\u0131 \u015f\u0259raitind\u0259. Waymo v\u0259 General Motors-dan bel\u0259 prototipl\u0259r var v\u0259 onlar bir ne\u00e7\u0259 \u015f\u0259h\u0259rd\u0259 i\u015f\u0259 salma\u011f\u0131 planla\u015fd\u0131r\u0131rlar v\u0259 real \u015f\u0259raitd\u0259 s\u0131naq ke\u00e7ir\u0259c\u0259kl\u0259r. Yandex-in Skolkovo v\u0259 Innopolis-d\u0259 avtonom taksi \u00fc\u00e7\u00fcn s\u0131naq b\u00f6lg\u0259l\u0259ri var: s\u00fcr\u00fc\u015f m\u00fch\u0259ndis t\u0259r\u0259find\u0259n s\u0259rni\u015fin oturaca\u011f\u0131nda oturark\u0259n m\u00fc\u015fahid\u0259 alt\u0131nda ba\u015f verir; ilin sonuna q\u0259d\u0259r \u015firk\u0259t avtonom ma\u015f\u0131nlar\u0131n say\u0131n\u0131 100-\u0259 q\u0259d\u0259r art\u0131rma\u011f\u0131 planla\u015fd\u0131r\u0131r.
    \nS\u0259viyy\u0259 5: Tam avtomatik s\u00fcr\u00fc\u015f, canl\u0131 s\u00fcr\u00fcc\u00fcn\u00fcn tam \u0259v\u0259zl\u0259nm\u0259si. Bel\u0259 sisteml\u0259r yoxdur v\u0259 onlar\u0131n yax\u0131n ill\u0259rd\u0259 g\u00f6r\u00fcnm\u0259si ehtimal\u0131 azd\u0131r.<\/p>\n

    B\u00fct\u00fcn bunlar\u0131 yax\u0131n zamanda g\u00f6rm\u0259k n\u0259 q\u0259d\u0259r reald\u0131r? Bu yerd\u0259 oxucunu m\u0259qal\u0259y\u0259 y\u00f6n\u0259ltm\u0259k ist\u0259rdik \u00ab2020-ci il\u0259 robotaksi ba\u015flatman\u0131n m\u00fcmk\u00fcn olmad\u0131\u011f\u0131n\u0131, Tesla-n\u0131n v\u0259d etdiyi kimi\u00bb<\/a><\/noindex>Bu, qism\u0259n 5G ba\u011flant\u0131s\u0131n\u0131n olmamas\u0131yla \u0259laq\u0259dard\u0131r: m\u00f6vcud 4G s\u00fcr\u0259tl\u0259ri yet\u0259rli deyil. H\u0259m\u00e7inin, avtonom avtomobill\u0259rin \u00e7ox y\u00fcks\u0259k qiym\u0259ti il\u0259 d\u0259 ba\u011fl\u0131d\u0131r: indiki vaxtda rentabelli deyill\u0259r, biznes modeli ayd\u0131n deyil. S\u00f6z\u00fcn q\u0131sas\u0131, m\u0259s\u0259l\u0259l\u0259r burada \"\u00e7\u0259tin\"dir, v\u0259 t\u0259sad\u00fcfi deyil ki, Gartner yaz\u0131r ki, 4 v\u0259 5-ci s\u0259viyy\u0259l\u0259rin k\u00fctl\u0259vi t\u0259tbiqinin proqnozu \u0259n az\u0131 10 ild\u0259n tez olmayacaq.<\/p>\n

    3.2. 3D g\u00f6rm\u0259 kameralar\u0131 (3D Sensing Cameras)<\/h2>\n

    \nS\u0259kkiz il \u0259vv\u0259l Microsoft-un Kinect oyun n\u0259zar\u0259t\u00e7isi 3D g\u00f6rm\u0259si \u00fc\u00e7\u00fcn \u0259l\u00e7atan v\u0259 nisb\u0259t\u0259n ucuz bir h\u0259ll t\u0259qdim ed\u0259r\u0259k s\u0259s-k\u00fcy yaratd\u0131. O vaxtdan b\u0259ri, Kinect il\u0259 idman-dans oyunlar\u0131 \u00f6z q\u0131sa y\u00fcks\u0259li\u015fini v\u0259 d\u00fc\u015f\u00fc\u015f\u00fcn\u00fc ya\u015fad\u0131, amma 3D kameralar s\u0259naye robotlar\u0131, pilotsuz avtomobill\u0259r v\u0259 mobil telefonlar t\u0259r\u0259find\u0259n \u00fcz tan\u0131ma \u00fc\u00e7\u00fcn istifad\u0259 edilm\u0259y\u0259 ba\u015fland\u0131. Texnologiya daha ucuz, daha kompakt v\u0259 daha \u0259l\u00e7atan oldu. <\/p>\n

    \"Gartner
    \nSamsung S10 telefonunda obyekt\u0259 m\u0259saf\u0259ni \u00f6l\u00e7\u0259n zaman u\u00e7u\u015fu (Time-of-Flight) kameras\u0131 var \u2013 fokuslanman\u0131 asanla\u015fd\u0131rmaq \u00fc\u00e7\u00fcn.     M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    \u018fg\u0259r bu m\u00f6vzu sizi maraqland\u0131r\u0131rsa, sizi d\u0259rinliyi \u00f6l\u00e7\u0259n kameralarla ba\u011fl\u0131 \u00e7ox yax\u015f\u0131 bir \u0259trafl\u0131 icmalara y\u00f6nl\u0259ndiririk: hiss\u0259 1<\/a><\/noindex>, hiss\u0259 2<\/a><\/noindex>.<\/p>\n

    3.3. Ki\u00e7ik y\u00fck da\u015f\u0131nmas\u0131 \u00fc\u00e7\u00fcn dronlar (Light Cargo Delivery Drones)<\/h2>\n

    \nBu il Amazon, 2 kg-a q\u0259d\u0259r ki\u00e7ik y\u00fck da\u015f\u0131ya bil\u0259n yeni u\u00e7an dronu s\u0259rgid\u0259 t\u0259qdim etdi. \u015e\u0259h\u0259r \u00fc\u00e7\u00fcn, onun t\u0131xaclar\u0131 il\u0259, bu m\u00fck\u0259mm\u0259l bir h\u0259ll kimi g\u00f6r\u00fcn\u00fcr. G\u0259lin g\u00f6r\u0259k, bu dronlar yax\u0131n g\u0259l\u0259c\u0259kd\u0259 nec\u0259 \u00f6z\u00fcn\u00fc g\u00f6st\u0259r\u0259c\u0259k. Burada ehtiyatl\u0131 \u015f\u00fcbh\u0259ni daxil etm\u0259k olar: dronun asanl\u0131qla o\u011furlanmas\u0131 sonra v\u0259 BPL-lara dair qanuni m\u0259hdudiyy\u0259tl\u0259r d\u0259 daxil olmaqla bir \u00e7ox problem var. Amazon Prime Air alt\u0131 il \u0259rzind\u0259 m\u00f6vcuddur, amma h\u0259l\u0259 d\u0259 test m\u0259rh\u0259l\u0259sind\u0259dir.<\/p>\n

    \"Gartner
    \nBu yazda t\u0259qdim olunan yeni Amazon dronu. \u0130\u00e7ind\u0259 \"Ulduz m\u00fcharib\u0259l\u0259ri\" elementl\u0259ri var.     M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    Amazon-dan ba\u015fqa, bu bazarda ba\u015fqa oyun\u00e7ular da var (daha \u0259trafl\u0131 icmal<\/a><\/noindex>), lakin he\u00e7 bir tamamlanm\u0131\u015f m\u0259hsul yoxdur: h\u0259r \u015fey \u2013 test m\u0259rh\u0259l\u0259sind\u0259 v\u0259 marketinq aksiyalar\u0131ndad\u0131r. Ayr\u0131-ayr\u0131l\u0131qda, Afrika m\u00fctl\u0259q maraql\u0131 ixtisasla\u015fm\u0131\u015f tibbi layih\u0259l\u0259rini qeyd etm\u0259k laz\u0131md\u0131r<\/a><\/noindex> : Qan donorlar\u0131n\u0131n da\u015f\u0131nmas\u0131 Gana (14,000 da\u015f\u0131nma, Zipline \u015firk\u0259ti) v\u0259 Ruanda (Matternet \u015firk\u0259ti) \u00fc\u00e7\u00fcn.<\/p>\n

    3.4. U\u00e7an avtonom n\u0259qliyyat (Flying Autonomous Vehicles)<\/h2>\n

    \nBurada m\u00fc\u0259yy\u0259n bir \u015fey dem\u0259k \u00e7\u0259tindir. Gartner-in fikrinc\u0259, bu, \u0259n az\u0131 10 ild\u0259n tez olmayacaq. \u00dcmumiyy\u0259tl\u0259, burada pilotsuz avtomobill\u0259rl\u0259 eyni probleml\u0259r var, lakin yeni bir \u00f6l\u00e7\u00fc \u2013 \u015faquli bir \u00f6l\u00e7\u00fc al\u0131r. U\u00e7an taksi qurma\u011f\u0131 planla\u015fd\u0131rd\u0131qlar\u0131n\u0131 Porsche, Boeing, Uber elan edibl\u0259r.<\/p>\n

    3.5. Art\u0131r\u0131lm\u0131\u015f reall\u0131q buludu (AR Cloud)<\/h2>\n

    \nGer\u00e7\u0259k d\u00fcnyan\u0131n davaml\u0131 r\u0259q\u0259msal sur\u0259ti, b\u00fct\u00fcn istifad\u0259\u00e7il\u0259r \u00fc\u00e7\u00fcn ortaq yeni reall\u0131q qat\u0131n\u0131 yaratma\u011fa imkan tan\u0131y\u0131r. Daha texniki dild\u0259 des\u0259k, burada a\u00e7\u0131q bulud platformas\u0131 yaratmaqdan dan\u0131\u015f\u0131l\u0131r ki, bu da inki\u015faf etdiricil\u0259rin \u00f6z AR t\u0259tbiql\u0259rini inteqrasiya etm\u0259sin\u0259 imkan versin. Monetizasiya modeli ayd\u0131nd\u0131r, bu, Steam-\u0131n bir analoqu kimidir. Fikir o q\u0259d\u0259r k\u00f6k salm\u0131\u015fd\u0131r ki, art\u0131q b\u0259zil\u0259ri AR-n\u0131n buludsuz sad\u0259c\u0259 faydas\u0131z oldu\u011funu d\u00fc\u015f\u00fcn\u00fcr.<\/p>\n

    G\u0259l\u0259c\u0259kd\u0259 bunun nec\u0259 g\u00f6r\u00fcn\u0259 bil\u0259c\u0259yin\u0259 dair q\u0131sa bir videoda t\u0259svir edilib. Bura, dig\u0259r \"Qara G\u00fczg\u00fc\" seriyalar\u0131ndan biri kimi g\u00f6r\u00fcn\u00fcr:
    \n

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

    Daha \u0259trafl\u0131 oxumaq olar m\u0259qal\u0259-d\u0259rcind\u0259.<\/a><\/noindex><\/p>\n

    4. \"Tamamlanm\u0131\u015f\" insan (Augmented Human)<\/h1>\n

    <\/p>\n

    4.1. Emosiyalar\u0131n a\u011f\u0131ll\u0131 g\u0259l\u0259c\u0259yi (Emotion AI)<\/h2>\n

    \n\u0130nsan emosiyalar\u0131n\u0131 nec\u0259 \u00f6l\u00e7m\u0259k, simulyasiya etm\u0259k v\u0259 cavab verm\u0259k olar? Burada m\u00fc\u015ft\u0259ril\u0259rd\u0259n biri Amazon Alexa kimi s\u0259sli assistentl\u0259ri istehsal ed\u0259n \u015firk\u0259tl\u0259rdir. Onlar evl\u0259rimizd\u0259 ger\u00e7\u0259kd\u0259n tam m\u0259m\u0259x olacaqlar, \u0259g\u0259r \u0259hval-ruhiyy\u0259ni tan\u0131ma\u011fa \u00f6yr\u0259n\u0259 bils\u0259l\u0259r: istifad\u0259\u00e7ini naraz\u0131 salan s\u0259b\u0259bi anlamaq, v\u0259ziyy\u0259ti d\u00fcz\u0259ltm\u0259y\u0259 \u00e7al\u0131\u015fmaq. \u00dcmumiyy\u0259tl\u0259, mesajdan daha \u00e7ox m\u0259lumat kontekstd\u0259dir. Kontekst is\u0259 \u00fcz ifad\u0259si, intonasiya v\u0259 qeyri-verbal davran\u0131\u015fd\u0131r.<\/p>\n

    Dig\u0259r praktik t\u0259tbiql\u0259rd\u0259n biri: i\u015f m\u00fcsahib\u0259si zaman\u0131 emosiyalar\u0131n analizi (video m\u00fcsahib\u0259l\u0259r), reklam \u00e7arxlar\u0131na v\u0259 ya dig\u0259r video m\u0259zmunlar\u0131na olan reaksiyan\u0131n qiym\u0259tl\u0259ndirilm\u0259si (g\u00fcl\u00fcms\u0259m\u0259l\u0259r, g\u00fcl\u00fc\u015fl\u0259r), \u00f6yr\u0259nm\u0259y\u0259 yard\u0131m (m\u0259s\u0259l\u0259n, ictimai \u00e7\u0131x\u0131\u015f bacar\u0131qlar\u0131 \u00fczr\u0259 m\u00fcst\u0259qil praktikalarda). <\/p>\n

    Bu m\u00f6vzuda 6 d\u0259qiq\u0259lik q\u0131sametrajl\u0131 film m\u00fc\u0259llifind\u0259n daha yax\u015f\u0131 he\u00e7 kim dan\u0131\u015fa bilm\u0259z Stealing Ur Feeling<\/a><\/noindex>. Z\u0259rif v\u0259 d\u0259bli \u00e7\u0259kilmi\u015f videoda, emosiyalar\u0131m\u0131z\u0131 marketinq m\u0259qs\u0259dl\u0259ri \u00fc\u00e7\u00fcn nec\u0259 \u00f6l\u00e7m\u0259k olaca\u011f\u0131n\u0131 g\u00f6st\u0259rir, v\u0259 \u00fcz\u00fcn\u00fcz\u00fcn anl\u0131q reaksiyalar\u0131ndan: pizza, itl\u0259r, Kanye West-i sevdiyiniz, h\u0259tta g\u0259lir s\u0259viyy\u0259niz v\u0259 t\u0259xmin\u0259n IQ-nuz bar\u0259d\u0259 m\u0259lumat \u0259ld\u0259 edilir. Yuxar\u0131dak\u0131 film sayt\u0131na daxil oldu\u011funuzda, notbukunuzun i\u00e7ind\u0259ki kameradan istifad\u0259 edil\u0259r\u0259k interaktiv video i\u015ftirak\u00e7\u0131s\u0131 olursunuz. Film art\u0131q bir ne\u00e7\u0259 film festival\u0131nda n\u00fcmayi\u015f etdirilib.<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    H\u0259tta bu maraql\u0131 t\u0259dqiqat var: m\u0259tnd\u0259 sarkazm\u0131 nec\u0259 a\u015fkar etm\u0259k olar. #sarcasm hashtag-\u0131 il\u0259 tvitl\u0259r g\u00f6t\u00fcr\u00fcl\u00fcb v\u0259 25,000 sarkazm olan tvitl\u0259rd\u0259n v\u0259 100,000 adi tvitd\u0259n ibar\u0259t \u00f6yr\u0259dici n\u00fcmun\u0259 yarad\u0131l\u0131b. TensorFlow kitabxanas\u0131 t\u0259tbiq edildi, sistem \u00f6yr\u0259dildi, n\u0259tic\u0259 budur:<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259<\/a><\/noindex><\/i><\/p>\n

    Bel\u0259likl\u0259, indi \u0259g\u0259r siz h\u0259mkar\u0131n\u0131z\u0131n v\u0259 ya dostunuzun dedikl\u0259rin\u0259 \u0259min deyilsinizs\u0259 \u2013 o siz\u0259 ciddi olaraq n\u0259s\u0259 dedi, yoxsa sarkazmla, - art\u0131q \u00f6yr\u0259dilmi\u015f neyron \u015f\u0259b\u0259k\u0259d\u0259n<\/a><\/noindex>! <\/p>\n

    4.2. Tamamlanm\u0131\u015f intellekt (Augmented Intelligence)<\/h2>\n

    \nMa\u015f\u0131n \u00f6yr\u0259nm\u0259 metodlar\u0131 il\u0259 intellektual i\u015fl\u0259rin avtomatla\u015fd\u0131r\u0131lmas\u0131. G\u00f6r\u00fcn\u00fcr, burada he\u00e7 bir yenilik yoxdur? Lakin burada formul\u0259nin \u00f6z\u00fc \u00f6n\u0259mlidir, x\u00fcsusil\u0259 d\u0259, bu, S\u00fcni \u0130ntellektl\u0259 (Artificial Intelligence) eyni q\u0131saltmaya malikdir. Bu, bizi \"g\u00fccl\u00fc\" v\u0259 \"z\u0259if\" AI-ya dair m\u00fcbahis\u0259l\u0259r\u0259 y\u00f6nl\u0259ndirir.
    \nG\u00fccl\u00fc AI, fantastik filml\u0259rd\u0259ki o c\u00fcrd\u00fcr ki, tamamil\u0259 insan a\u011fl\u0131n\u0131n ekvivalentidir v\u0259 \u00f6z\u00fcn\u00fc \u015f\u0259xsiyy\u0259t kimi ba\u015fa d\u00fc\u015f\u00fcr. Bel\u0259 bir \u015fey h\u0259l\u0259 m\u00f6vcud deyil v\u0259 \u00fcmumiyy\u0259tl\u0259, bel\u0259 bir \u015feyin olub-olmayaca\u011f\u0131 da m\u0259lum deyil.<\/p>\n

    Z\u0259if AI m\u00fcst\u0259qil bir \u015f\u0259xsiyy\u0259t deyil, sad\u0259c\u0259 insan\u0131n k\u00f6m\u0259k\u00e7isi-dir. O, insan\u0131n b\u0259nz\u0259rd\u0259n d\u00fc\u015f\u00fcnm\u0259 iddias\u0131nda deyil, sad\u0259c\u0259 informasiya m\u0259s\u0259l\u0259l\u0259rini h\u0259ll ed\u0259 bilir, m\u0259s\u0259l\u0259n, \u015f\u0259kilin n\u0259yi \u0259ks etdiyini t\u0259yin etm\u0259k v\u0259 ya m\u0259tni t\u0259rc\u00fcm\u0259 etm\u0259k.<\/p>\n

    \"Gartner
    \n    M\u0259nb\u0259 <\/a><\/noindex><\/i><\/p>\n

    Bu m\u0259nada Art\u0131r\u0131lm\u0131\u015f Z\u0259ka tamamil\u0259 \"z\u0259if AI\"-dir v\u0259 bu ifad\u0259 \u00f6zl\u00fcy\u00fcnd\u0259 uy\u011fundur, \u00e7\u00fcnki burada ham\u0131n\u0131n arzulad\u0131\u011f\u0131 (ya da \"ma\u015f\u0131nlar\u0131n \u00fcsyan\u0131\" haqq\u0131nda \u00e7oxsayl\u0131 m\u00fczakir\u0259l\u0259ri xat\u0131rlayanda) o \"g\u00fccl\u00fc AI\"-ni g\u00f6rm\u0259k \u00fc\u00e7\u00fcn bir qar\u0131\u015f\u0131ql\u0131q yaratm\u0131r. Art\u0131r\u0131lm\u0131\u015f Z\u0259ka ifad\u0259si il\u0259, sanki biz f\u0259rqli bir filmin q\u0259hr\u0259manlar\u0131 oluruk: elmi fantastikadan (m\u0259s\u0259l\u0259n, Azimovun \"M\u0259n, robot\" roman\u0131) kiberpank d\u00fcnyas\u0131na ke\u00e7irik (bu janrdak\u0131 \"a\u011f\u0131rl\u0131qlar\" insan\u0131n imkanlar\u0131n\u0131 art\u0131ran m\u00fcxt\u0259lif implantlar\u0131 ifad\u0259 edir).<\/p>\n

    Nec\u0259 dedil\u0259r<\/a><\/noindex> Erik Brynjolfsson v\u0259 Andrew McAfee: \"G\u0259l\u0259c\u0259k 10 il \u0259rzind\u0259 bel\u0259 olacaq. AI menecerl\u0259ri \u0259v\u0259z etmir, AI istifad\u0259 ed\u0259n menecerl\u0259r h\u0259l\u0259 istifad\u0259 etm\u0259y\u0259nl\u0259ri \u0259v\u0259z ed\u0259c\u0259k\"<\/p>\n

    N\u00fcmun\u0259l\u0259r:<\/p>\n