{"id":28,"date":"2025-12-18T22:29:29","date_gmt":"2025-12-18T19:29:29","guid":{"rendered":"https:\/\/sultansalblog.com.tr\/?p=28"},"modified":"2025-12-18T23:30:09","modified_gmt":"2025-12-18T20:30:09","slug":"what-ive-learned-from-road-trips","status":"publish","type":"post","link":"https:\/\/sultansalblog.com.tr\/index.php\/2025\/12\/18\/what-ive-learned-from-road-trips\/","title":{"rendered":"\ud83e\udde0 Yapay Zek\u00e2 Nas\u0131l \u00c7al\u0131\u015f\u0131r? Ad\u0131m Ad\u0131m Basit Anlat\u0131m"},"content":{"rendered":"

\ud83e\udde0 Yapay Zek\u00e2 Nas\u0131l \u00c7al\u0131\u015f\u0131r? Ad\u0131m Ad\u0131m Basit Anlat\u0131m<\/strong><\/h1>\n

\u0130lk yaz\u0131m\u0131zda \u201cYapay Zek\u00e2 nedir?\u201d ve t\u00fcrlerini inceledik.
\n\u015eimdi s\u0131ra, bu zek\u00e2n\u0131n bir bilgisayar\u0131n nas\u0131l olup da bir kediyi tan\u0131yabildi\u011fini, e-postalar\u0131n spam olup olmad\u0131\u011f\u0131n\u0131 ay\u0131rt edebildi\u011fini veya sana izleyebilece\u011fin bir film \u00f6nerebildi\u011fini ad\u0131m ad\u0131m g\u00f6relim.<\/p>\n

Bu, sihir de\u011fil \u2014 tamamen veri, algoritma ve \u00f6\u011frenme<\/strong> \u00fczerine kurulu g\u00fc\u00e7l\u00fc bir sistemdir.
\nGelin \u015fimdi bu s\u00fcreci sadele\u015ftirip anlayal\u0131m.<\/p>\n

\ud83d\udca1 Yapay Zek\u00e2n\u0131n Arkas\u0131ndaki Mant\u0131k<\/strong><\/h2>\n

Yapay zek\u00e2 (YZ) kula\u011fa karma\u015f\u0131k gelse de temelde yapt\u0131\u011f\u0131 \u015fey olduk\u00e7a basit:
\nVeri toplamak, veriden \u00f6\u011frenmek ve \u00f6\u011frendi\u011fi bilgiyi kullanarak karar vermek.<\/strong><\/p>\n

Bir insan\u0131n deneyimlerinden \u00f6\u011frendi\u011fi gibi, YZ de verilerden \u00f6\u011frenir.<\/strong>
\nFark\u0131, duygular yerine matematiksel algoritmalarla<\/strong> \u00e7al\u0131\u015fmas\u0131d\u0131r.<\/p>\n

\ud83d\udd0d 1. Ad\u0131m: Veri (YZ\u2019nin Ham Maddesi)<\/strong><\/h4>\n

Yapay zek\u00e2n\u0131n yak\u0131t\u0131 veri<\/strong>dir.
\nNe kadar fazla ve \u00e7e\u015fitli veriyle beslenirse, o kadar do\u011fru sonu\u00e7 verir.<\/p>\n

\ud83e\udde0 Veri t\u00fcrleri:<\/strong><\/p>\n

    \n
  • Foto\u011fraflar, ses kay\u0131tlar\u0131, yaz\u0131lar, say\u0131sal tablolar, davran\u0131\u015f kay\u0131tlar\u0131.<\/li>\n
  • \u00d6rne\u011fin bir y\u00fcz tan\u0131ma sistemi binlerce farkl\u0131 y\u00fcz foto\u011fraf\u0131yla e\u011fitilir.<\/li>\n
  • Bir \u00e7eviri uygulamas\u0131 milyonlarca c\u00fcmle \u00f6rne\u011fini inceleyerek \u00f6\u011frenir.<\/li>\n<\/ul>\n

    Bu veriler etiketlenir<\/strong>, yani sistemin \u00f6\u011frenebilmesi i\u00e7in i\u015faretlenir:<\/p>\n

    \u201cBu bir kedi\u201d, \u201cBu bir k\u00f6pek\u201d, \u201cBu spam\u201d, \u201cBu de\u011fil.\u201d<\/p>\n

    Veri olmazsa, yapay zek\u00e2 da olmaz. Bu y\u00fczden teknoloji d\u00fcnyas\u0131nda s\u0131k\u00e7a \u201cVeri yeni alt\u0131nd\u0131r<\/strong>\u201d denir.<\/p>\n

    \u2699\ufe0f 2. Ad\u0131m: Algoritma (YZ\u2019nin Motoru)<\/strong><\/h4>\n

    Veri topland\u0131ktan sonra, onu anlamland\u0131racak bir algoritma<\/strong> devreye girer.
    \nAlgoritma<\/strong>, bir problemi \u00e7\u00f6zmek i\u00e7in izlenen ad\u0131mlar dizisidir \u2014 yani YZ\u2019nin \u201cre\u00e7etesi\u201ddir.<\/p>\n

    \ud83e\udde9 \u00d6rnek:
    \n\u201cE\u011fer bu foto\u011frafta iki g\u00f6z, bir burun ve bir a\u011f\u0131z varsa \u2192 bu bir insan y\u00fcz\u00fcd\u00fcr.\u201d<\/p>\n

    Bu mant\u0131k, algoritmalar sayesinde sistematik bi\u00e7imde i\u015fler.
    \nAma as\u0131l b\u00fcy\u00fc, algoritman\u0131n sadece kurallar\u0131 uygulamakla kalmay\u0131p \u00f6\u011frenmeye ba\u015flamas\u0131yla<\/strong> ger\u00e7ekle\u015fir.<\/p>\n

    \ud83e\udd16 3. Ad\u0131m: Makine \u00d6\u011frenimi (Machine Learning)<\/strong><\/h4>\n

    Makine \u00d6\u011frenimi (ML)<\/strong>, Yapay Zek\u00e2\u2019n\u0131n kendi kendine \u00f6\u011frenmesini sa\u011flar.
    \nYani makine, a\u00e7\u0131k\u00e7a programlanmadan deneyimden \u00f6\u011frenir.<\/strong><\/p>\n

    \u00d6rnekle a\u00e7\u0131klayal\u0131m:<\/p>\n

      \n
    1. Sisteme binlerce \u201cspam\u201d ve \u201cspam olmayan\u201d e-posta verilir.<\/li>\n
    2. Algoritma, ortak kelime ve desenleri bulur.<\/li>\n
    3. Art\u0131k yeni bir e-posta geldi\u011finde \u201cBu spam mi de\u011fil mi?\u201d tahmin edebilir.<\/li>\n<\/ol>\n

      Bu s\u00fcre\u00e7 sonunda ortaya bir model<\/strong> \u00e7\u0131kar.
      \nModel, \u00f6\u011frendiklerini kullanarak yeni verilerde tahmin yapabilir.<\/strong><\/p>\n

      K\u0131sacas\u0131, makine \u00f6\u011frenmesi bir sistemin \u201cdeneyim kazand\u0131k\u00e7a ak\u0131llanmas\u0131d\u0131r.\u201d<\/p>\n

      \ud83e\udde0 4. Ad\u0131m: Derin \u00d6\u011frenme (Deep Learning)<\/strong><\/h4>\n

      Makine \u00d6\u011frenimi\u2019nin bir \u00fcst seviyesi olan Derin \u00d6\u011frenme (DL)<\/strong>, g\u00fcn\u00fcm\u00fcz\u00fcn en g\u00fc\u00e7l\u00fc yapay zek\u00e2 sistemlerinin temelidir.<\/p>\n

      DL\u2019nin s\u0131rr\u0131, Yapay Sinir A\u011flar\u0131 (Neural Networks)<\/strong> kullanmas\u0131d\u0131r \u2014 yani insan beynini taklit eden katmanl\u0131 yap\u0131lar.<\/p>\n

      Bu a\u011flarda bilgi, bir katmandan di\u011ferine ge\u00e7erken daha karma\u015f\u0131k hale gelir:<\/p>\n\n\n\n\n\n\n\n\n
      Katman<\/strong><\/td>\nNe \u00d6\u011frenir?<\/strong><\/td>\n<\/tr>\n<\/thead>\n
      1. Katman<\/td>\nBasit \u00e7izgiler, kenarlar<\/td>\n<\/tr>\n
      2. Katman<\/td>\n\u015eekiller, renk ge\u00e7i\u015fleri<\/td>\n<\/tr>\n
      3. Katman<\/td>\nNesne b\u00f6l\u00fcmleri (g\u00f6z, kulak, tekerlek)<\/td>\n<\/tr>\n
      4. Katman<\/td>\nTam nesneler (\u201ckedi\u201d, \u201caraba\u201d)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      \ud83d\udcf8 \u00d6rnek:<\/strong>
      \nBir resim tan\u0131ma sistemi, kediyi tan\u0131mak i\u00e7in \u00f6nce \u015fekilleri, sonra t\u00fcy desenlerini analiz eder ve en sonunda \u201cevet, bu bir kedi\u201d der.<\/p>\n

      Bu nedenle s\u00fcr\u00fcc\u00fcs\u00fcz ara\u00e7lar, otomatik \u00e7eviri sistemleri ve ChatGPT gibi dil modelleri<\/strong> derin \u00f6\u011frenme sayesinde \u00e7al\u0131\u015f\u0131r.<\/p>\n

      \ud83d\udd01 5. Ad\u0131m: S\u00fcrekli \u00d6\u011frenme ve \u0130yile\u015fme<\/strong><\/h4>\n

      YZ sistemleri bir kez \u00f6\u011frenip durmaz.
      \nYeni veri geldik\u00e7e kendilerini g\u00fcnceller<\/strong> ve daha do\u011fru sonu\u00e7lar \u00fcretir.<\/p>\n

      \u00d6rne\u011fin:<\/p>\n

        \n
      • Google Translate her g\u00fcn yeni \u00e7eviri \u00f6rneklerinden \u00f6\u011frenir.<\/li>\n
      • Spotify, dinleme al\u0131\u015fkanl\u0131klar\u0131n\u0131 analiz ederek \u00f6nerilerini geli\u015ftirir.<\/li>\n
      • ChatGPT, her etkile\u015fimle dil kal\u0131plar\u0131n\u0131 daha iyi anlar.<\/li>\n<\/ul>\n

        Yani yapay zek\u00e2 ya\u015fland\u0131k\u00e7a de\u011fil, \u00f6\u011frendik\u00e7e geli\u015fir.<\/strong><\/p>\n

        \ud83d\ude80 6. Ad\u0131m: Sonu\u00e7 (Tahmin ve Karar)<\/strong><\/h4>\n

        Model art\u0131k haz\u0131rd\u0131r.
        \nYeni bir veri geldi\u011finde (foto\u011fraf, e-posta veya ses komutu), sistem \u00f6\u011frendiklerini kullanarak h\u0131zl\u0131ca karar verir:<\/p>\n

          \n
        • \u201cBu bir kedidir.\u201d<\/li>\n
        • \u201cBu e-posta spam de\u011fildir.\u201d<\/li>\n
        • \u201cBu c\u00fcmle \u0130ngilizceden T\u00fcrk\u00e7eye \u015f\u00f6yle \u00e7evrilir.\u201d<\/li>\n<\/ul>\n

          Ve her karar, sistemin kendini bir ad\u0131m daha geli\u015ftirmesine<\/strong> katk\u0131 sa\u011flar.<\/p>\n

          \ud83e\udde9 \u00d6zet: Yapay Zek\u00e2n\u0131n \u00c7al\u0131\u015fma Zinciri<\/strong><\/h4>\n

          Veri \u2192 Algoritma \u2192 \u00d6\u011frenme \u2192 Tahmin \/ Karar<\/strong><\/p>\n

          Yani YZ, bir insan gibi deneyim kazan\u0131r \u2014
          \ntek fark\u0131, bunu milyonlarca kez daha h\u0131zl\u0131 yapmas\u0131d\u0131r.<\/p>\n

          \ud83c\udf0d G\u00fcnl\u00fck Hayattan Basit Bir \u00d6rnek<\/strong><\/p>\n

          Telefonundaki kamera \u201cg\u00fcl\u00fcmseme alg\u0131lama\u201d \u00f6zelli\u011fini d\u00fc\u015f\u00fcn:<\/p>\n

            \n
          1. Kamera g\u00f6r\u00fcnt\u00fcy\u00fc al\u0131r (veri<\/strong>).<\/li>\n
          2. Sistemdeki kurallar uygulan\u0131r (algoritma<\/strong>).<\/li>\n
          3. Sistem ge\u00e7mi\u015fteki y\u00fcz ifadelerinden \u00f6\u011frenmi\u015ftir (makine \u00f6\u011frenmesi<\/strong>).<\/li>\n
          4. G\u00fcl\u00fcmsemeyi tespit eder ve foto\u011fraf\u0131 \u00e7eker (karar<\/strong>).<\/li>\n<\/ol>\n

            Yani asl\u0131nda g\u00f6rd\u00fc\u011f\u00fcn \u201csihir\u201d, sadece \u00e7ok h\u0131zl\u0131 \u00e7al\u0131\u015fan bir \u00f6\u011frenme d\u00f6ng\u00fcs\u00fcd\u00fcr.<\/p>\n

            \"\"\ud83d\udd2e Sonu\u00e7: Makinelerin \u00d6\u011frenme G\u00fcc\u00fc<\/strong><\/h3>\n

            Bug\u00fcn yaz\u0131 yazan, konu\u015fan, resim tan\u0131yan hatta m\u00fczik besteleyen yapay zek\u00e2 sistemleri, bu ad\u0131mlar\u0131n \u00fcr\u00fcn\u00fcd\u00fcr.
            \nYZ art\u0131k gelece\u011fin de\u011fil, bug\u00fcn\u00fcn teknolojisidir.<\/p>\n

            \u201cYapay zek\u00e2 ak\u0131ll\u0131 g\u00f6r\u00fcn\u00fcr, \u00e7\u00fcnk\u00fc onu \u00f6\u011freten insanlar zekidir.\u201d \ud83d\udca1<\/p>\n","protected":false},"excerpt":{"rendered":"

            \ud83e\udde0 Yapay Zek\u00e2 Nas\u0131l \u00c7al\u0131\u015f\u0131r? Ad\u0131m Ad\u0131m Basit Anlat\u0131m \u0130lk yaz\u0131m\u0131zda \u201cYapay Zek\u00e2 nedir?\u201d ve t\u00fcrlerini inceledik. \u015eimdi s\u0131ra, bu zek\u00e2n\u0131n bir bilgisayar\u0131n nas\u0131l olup da bir kediyi tan\u0131yabildi\u011fini, e-postalar\u0131n spam olup olmad\u0131\u011f\u0131n\u0131 ay\u0131rt edebildi\u011fini veya sana izleyebilece\u011fin bir film \u00f6nerebildi\u011fini ad\u0131m ad\u0131m g\u00f6relim. Bu, sihir de\u011fil \u2014 tamamen veri, algoritma ve \u00f6\u011frenme \u00fczerine kurulu…<\/p>\n","protected":false},"author":1,"featured_media":56,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_eb_attr":"","footnotes":""},"categories":[8],"tags":[5],"class_list":["post-28","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-yazi","tag-yapayzeka-makineogrenmesi-derinogrenme-teknoloji-aiteknolojileri-dijitalgelecek"],"_links":{"self":[{"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/posts\/28","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/comments?post=28"}],"version-history":[{"count":5,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/posts\/28\/revisions"}],"predecessor-version":[{"id":72,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/posts\/28\/revisions\/72"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/media\/56"}],"wp:attachment":[{"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/media?parent=28"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/categories?post=28"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/sultansalblog.com.tr\/index.php\/wp-json\/wp\/v2\/tags?post=28"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}