{"id":684,"date":"2026-08-17T09:28:23","date_gmt":"2026-08-17T12:28:23","guid":{"rendered":"https:\/\/myozden.info\/?p=684"},"modified":"2026-08-17T09:33:09","modified_gmt":"2026-08-17T12:33:09","slug":"pbim-aos-agentic-yapay-zeka","status":"publish","type":"post","link":"https:\/\/myozden.info\/index.php\/2026\/08\/17\/pbim-aos-agentic-yapay-zeka\/","title":{"rendered":"PBIM-AOS: Alg\u0131sal Darbo\u011fazlardan Agentic Yapay Zek\u00e2 ile Pedagojik Karar Deste\u011fine"},"content":{"rendered":"\n

PBIM-AOS (Perceptual Bottleneck Inventory<\/a> \u2013 Agentic Operating System), \u00f6\u011frenci hatalar\u0131n\u0131n arkas\u0131ndaki alg\u0131sal darbo\u011fazlar\u0131 kan\u0131ta dayal\u0131 olarak incelemek ve \u00f6\u011fretmene pedagojik karar deste\u011fi sa\u011flamak amac\u0131yla geli\u015ftirilmekte olan agentic yapay zek\u00e2 yakla\u015f\u0131m\u0131d\u0131r.<\/p>\n\n\n\n

\u00d6\u011frenci hatas\u0131n\u0131 g\u00f6rmek yeterli mi, yoksa yapay zek\u00e2 bu hatan\u0131n arkas\u0131ndaki \u00f6\u011frenme sorununu anlamada \u00f6\u011fretmene yard\u0131mc\u0131 olabilir mi?<\/h3>\n\n\n\n

Yapay zek\u00e2n\u0131n e\u011fitimdeki ilk yayg\u0131n kullan\u0131m bi\u00e7imlerinden biri olduk\u00e7a basitti:<\/p>\n\n\n\n

\u00d6\u011frenci \u2192 Yapay zek\u00e2 \u2192 Cevap<\/strong><\/p>\n\n\n\n

\u00d6\u011frenci bir soru soruyor, yapay zek\u00e2 ise bildiklerine ve verilen ba\u011flama dayanarak bir cevap \u00fcretiyordu.<\/p>\n\n\n\n

\u00dcretken yapay zek\u00e2n\u0131n e\u011fitimde kullan\u0131m\u0131 a\u00e7\u0131s\u0131ndan bu \u00f6nemli bir geli\u015fmeydi. Ancak \u00f6\u011fretme ve \u00f6\u011frenme s\u00fcreci yaln\u0131zca soru ve cevaplardan olu\u015fmaz.<\/p>\n\n\n\n

Bu yakla\u015f\u0131m, yapay zek\u00e2n\u0131n pedagojik kararlar\u0131n yerine ge\u00e7mesi yerine \u00f6\u011fretmenin mesleki yarg\u0131s\u0131n\u0131 desteklemesi gerekti\u011fini vurgulayan insan-merkezli yapay zek\u00e2 yakla\u015f\u0131mlar\u0131<\/a>yla da uyumludur.<\/p>\n\n\n\n

Bir \u00f6\u011frenci matematik problemini yanl\u0131\u015f \u00e7\u00f6zd\u00fc\u011f\u00fcnde \u00f6\u011fretmen a\u00e7\u0131s\u0131ndan as\u0131l \u00f6nemli soru \u00e7o\u011fu zaman:<\/p>\n\n\n\n

\u201cDo\u011fru cevap nedir?\u201d<\/strong><\/p>\n\n\n\n

de\u011fil,<\/p>\n\n\n\n

\u201c\u00d6\u011frenci neden b\u00f6yle d\u00fc\u015f\u00fcnd\u00fc?\u201d<\/strong><\/p>\n\n\n\n

sorusudur.<\/p>\n\n\n\n

Perceptual Bottleneck Inventory (PBI) yakla\u015f\u0131m\u0131n\u0131n \u00e7\u0131k\u0131\u015f noktalar\u0131ndan biri de bu ayr\u0131md\u0131r. \u00d6\u011frencinin hatas\u0131 bize neyin yanl\u0131\u015f gitti\u011fini<\/em> g\u00f6sterebilir; alg\u0131sal darbo\u011faz ise neden yanl\u0131\u015f gitti\u011fini<\/em> anlamam\u0131za yard\u0131mc\u0131 olabilir.<\/p>\n\n\n\n

Bu d\u00fc\u015f\u00fcncenin matematik e\u011fitimindeki sistematik uygulamas\u0131 olan PBIM \u2014 Perceptual Bottleneck Inventory for Mathematics<\/strong>, \u00f6\u011frenme \u00e7\u0131kt\u0131lar\u0131, matematiksel kavramlar, tipik \u00f6\u011frenci hatalar\u0131, alg\u0131sal darbo\u011fazlar, tan\u0131lama sorular\u0131 ve \u00f6\u011fretim m\u00fcdahaleleri aras\u0131nda yap\u0131land\u0131r\u0131lm\u0131\u015f ili\u015fkiler kurmaktad\u0131r.<\/p>\n\n\n\n

Fakat burada yeni bir soru ortaya \u00e7\u0131kmaktad\u0131r:<\/p>\n\n\n\n

\n

Bu pedagojik bilgi yaln\u0131zca bir envanterde saklanmak yerine, \u00f6\u011frencinin \u00f6\u011frenme s\u00fcrecinde aktif olarak kullan\u0131labilir mi?<\/strong><\/p>\n<\/blockquote>\n\n\n\n

PBIM-AOS \u2014 PBIM Agentic Operating System<\/strong> \u2014 bu soruya yan\u0131t arayan bir sonraki ad\u0131md\u0131r.<\/p>\n\n\n\n

\n\n\n\n

Chatbot’tan AI ajanlar\u0131na<\/h2>\n\n\n\n

Bir chatbot ile agentic bir yapay zek\u00e2 sistemi aras\u0131nda \u00f6nemli bir fark vard\u0131r.<\/p>\n\n\n\n

Bir chatbotu \u00e7ok bilgili tek bir \u00f6\u011fretmene benzetebiliriz.<\/p>\n\n\n\n

Agentic Operating System ise farkl\u0131 uzmanl\u0131klara sahip \u00f6\u011fretmenleri, \u00f6l\u00e7me-de\u011ferlendirme uzmanlar\u0131n\u0131 ve rehberlik s\u00fcre\u00e7lerini koordine eden bir okul y\u00f6netimine benzetilebilir. Kaynak belgede bu ayr\u0131m, Agentic OS’nin planlama, karar verme, g\u00f6rev payla\u015f\u0131m\u0131, ara\u00e7 kullan\u0131m\u0131 ve sonu\u00e7lar\u0131n de\u011ferlendirilmesini y\u00f6neten bir \u00fcst katman olmas\u0131yla a\u00e7\u0131klanmaktad\u0131r.<\/p>\n\n\n\n

Bu nedenle agentic yakla\u015f\u0131mda yapay zek\u00e2 art\u0131k yaln\u0131zca cevap \u00fcretmez.<\/p>\n\n\n\n

Bir hedef do\u011frultusunda farkl\u0131 g\u00f6revleri yerine getiren uzman ajanlar birlikte \u00e7al\u0131\u015fabilir.<\/p>\n\n\n\n

E\u011fitim a\u00e7\u0131s\u0131ndan d\u00fc\u015f\u00fcn\u00fcld\u00fc\u011f\u00fcnde s\u00fcre\u00e7 \u00f6rne\u011fin \u015f\u00f6yle olabilir:<\/p>\n\n\n\n

\u00d6\u011frenci yan\u0131t\u0131
\u2193
Tan\u0131lama
\u2193
Alg\u0131sal darbo\u011faz\u0131n belirlenmesi
\u2193
Kan\u0131tlar\u0131n de\u011ferlendirilmesi
\u2193
\u00d6\u011fretim m\u00fcdahalesi
\u2193
\u00d6\u011frenme etkinli\u011fi
\u2193
De\u011ferlendirme
\u2193
Yeniden tan\u0131lama<\/strong><\/p>\n\n\n\n

Bu yap\u0131 tek seferlik bir yapay zek\u00e2 cevab\u0131ndan farkl\u0131 olarak s\u00fcrekli bir \u00f6\u011frenme d\u00f6ng\u00fcs\u00fc<\/strong> olu\u015fturur. PBIM-AOS’un \u00f6ng\u00f6rd\u00fc\u011f\u00fc temel d\u00f6ng\u00fc de \u00f6\u011frenci yan\u0131t\u0131ndan ba\u015flay\u0131p tan\u0131lama, darbo\u011faz, m\u00fcdahale ve de\u011ferlendirme \u00fczerinden yeniden tan\u0131lamaya ula\u015fmaktad\u0131r.<\/p>\n\n\n\n

\n\n\n\n

PBIM’den PBIM-AOS’a<\/h1>\n\n\n\n

PBIM ile PBIM-AOS aras\u0131ndaki ili\u015fkiyi belki de en k\u0131sa \u015fekilde \u015f\u00f6yle ifade edebiliriz:<\/p>\n\n\n\n

\n

PBIM \u201cNe biliyoruz?\u201d sorusuna cevap verir.
PBIM-AOS ise \u201cBu bilgiyi \u00f6\u011frenci i\u00e7in nas\u0131l i\u015fletebiliriz?\u201d sorusuna cevap arar.<\/strong><\/p>\n<\/blockquote>\n\n\n\n

PBIM bir pedagojik bilgi katman\u0131d\u0131r<\/strong>.<\/p>\n\n\n\n

Bir \u00f6\u011frenme \u00e7\u0131kt\u0131s\u0131na ili\u015fkin matematiksel kavramlar\u0131, tipik \u00f6\u011frenci hatalar\u0131n\u0131, olas\u0131 alg\u0131sal darbo\u011fazlar\u0131, tan\u0131lama sorular\u0131n\u0131 ve \u00f6\u011fretim m\u00fcdahalelerini yap\u0131land\u0131r\u0131r.<\/p>\n\n\n\n

PBIM-AOS ise bu bilgiyi kullanan pedagojik y\u00fcr\u00fctme ve karar katman\u0131d\u0131r<\/strong>. Kaynak belgede bu iki katman a\u00e7\u0131k bi\u00e7imde \u201cPBIM = pedagojik bilgi katman\u0131\u201d ve \u201cPBIM-AOS = pedagojik y\u00fcr\u00fctme ve karar katman\u0131\u201d olarak ayr\u0131lmaktad\u0131r.<\/p>\n\n\n\n

Bu nedenle PBIM-AOS, PBIM’in yerine ge\u00e7mez.<\/p>\n\n\n\n

PBIM’in \u00fczerine kurulur.<\/strong><\/p>\n\n\n\n

\n\n\n\n

PBIM-AOS’un uzman ajanlar\u0131<\/h2>\n\n\n\n

PBIM-AOS mimarisinde pedagojik s\u00fcrecin farkl\u0131 b\u00f6l\u00fcmlerini ele alan uzman ajanlar \u00f6ng\u00f6r\u00fclmektedir.<\/p>\n\n\n\n

Outcome Agent<\/strong>, \u00f6\u011frencinin \u00e7al\u0131\u015ft\u0131\u011f\u0131 \u00f6\u011frenme \u00e7\u0131kt\u0131s\u0131n\u0131 belirler.<\/p>\n\n\n\n

Bottleneck Agent<\/strong>, PBIM envanterindeki bilgilerden yararlanarak olas\u0131 alg\u0131sal darbo\u011fazlar\u0131 de\u011ferlendirir.<\/p>\n\n\n\n

Misconception Agent<\/strong>, \u00f6\u011frenci yan\u0131t\u0131n\u0131 PBIM’de tan\u0131mlanm\u0131\u015f tipik \u00f6\u011frenci hatalar\u0131yla kar\u015f\u0131la\u015ft\u0131r\u0131r.<\/p>\n\n\n\n

Evidence Agent<\/strong>, \u00f6\u011frencinin cevab\u0131ndan tan\u0131sal de\u011feri olan kan\u0131tlar\u0131 \u00e7\u0131kar\u0131r.<\/p>\n\n\n\n

Intervention Agent<\/strong>, belirlenen sorun ve mevcut kan\u0131tlara g\u00f6re uygun \u00f6\u011fretim m\u00fcdahalelerini \u00f6nerir.<\/p>\n\n\n\n

Activity Agent<\/strong>, \u00f6\u011frenci veya s\u0131n\u0131f i\u00e7in yapay zek\u00e2 destekli \u00f6\u011frenme etkinlikleri olu\u015fturur.<\/p>\n\n\n\n

Assessment Agent<\/strong>, tan\u0131y\u0131 do\u011frulamaya ve m\u00fcdahale sonras\u0131nda \u00f6\u011frenmeyi yeniden de\u011ferlendirmeye y\u00f6nelik sorular \u00fcretir.<\/p>\n\n\n\n

Teacher Coach Agent<\/strong> ise \u00f6\u011fretmene pedagojik a\u00e7\u0131klamalar, alternatif stratejiler ve s\u0131n\u0131f i\u00e7i uygulama \u00f6nerileri sunar.<\/p>\n\n\n\n

Bu sekiz uzman ajan PBIM-AOS’un ba\u015flang\u0131\u00e7 mimarisinin temelini olu\u015fturmaktad\u0131r.<\/p>\n\n\n\n

Ancak ajanlar\u0131n birbirinden ba\u011f\u0131ms\u0131z \u00e7al\u0131\u015fmas\u0131 yeterli de\u011fildir.<\/p>\n\n\n\n

Burada PBIM-AOS’un merkezi bile\u015fenlerinden biri devreye girer:<\/p>\n\n\n\n

Orchestrator<\/h2>\n\n\n\n

Orchestrator’\u0131 sistemin orkestra \u015fefi<\/strong> gibi d\u00fc\u015f\u00fcnebiliriz.<\/p>\n\n\n\n

Orchestrator \u015fu sorularla ilgilenir:<\/p>\n\n\n\n

Hangi ajan \u00e7al\u0131\u015fmal\u0131?
Bir sonraki ajan hangisi olmal\u0131?
Yeterli kan\u0131t var m\u0131?
Ek bir tan\u0131lama sorusu gerekli mi?
Tan\u0131 hipotezi yeterince g\u00fc\u00e7l\u00fc m\u00fc?
\u00d6\u011fretmen onay\u0131 gerekiyor mu?
M\u00fcdahaleye ge\u00e7ilebilir mi?<\/strong><\/p>\n\n\n\n

Dolay\u0131s\u0131yla agentic sistem yaln\u0131zca bir\u00e7ok yapay zek\u00e2 modelinin ayn\u0131 anda kullan\u0131lmas\u0131 anlam\u0131na gelmez.<\/p>\n\n\n\n

As\u0131l \u00f6zellik, uzman bile\u015fenlerin ortak bir pedagojik ama\u00e7 do\u011frultusunda koordine edilmesidir<\/strong>. Kaynakta Orchestrator’\u0131n hangi ajan\u0131n \u00e7al\u0131\u015faca\u011f\u0131na, ek kan\u0131t gerekip gerekmedi\u011fine ve \u00f6\u011fretmen onay\u0131na ihtiya\u00e7 olup olmad\u0131\u011f\u0131na karar veren merkezi bile\u015fen oldu\u011fu belirtilmektedir.<\/p>\n\n\n\n

\n\n\n\n

Kesin tan\u0131 de\u011fil, kan\u0131ta dayal\u0131 hipotez<\/h1>\n\n\n\n

PBIM-AOS tasar\u0131m\u0131n\u0131n pedagojik a\u00e7\u0131dan en \u00f6nemli ilkelerinden biri burada ortaya \u00e7\u0131kmaktad\u0131r.<\/p>\n\n\n\n

Bir \u00f6\u011frencinin tek bir yanl\u0131\u015f cevab\u0131ndan hareketle:<\/p>\n\n\n\n

\u201cBu \u00f6\u011frencide \u015fu alg\u0131sal darbo\u011faz vard\u0131r.\u201d<\/strong><\/p>\n\n\n\n

demek do\u011fru olmayabilir.<\/p>\n\n\n\n

Bunun yerine sistemin \u015f\u00f6yle d\u00fc\u015f\u00fcnmesi gerekir:<\/p>\n\n\n\n

Olas\u0131 darbo\u011faz
+
\u00d6\u011frenci yan\u0131t\u0131ndan elde edilen kan\u0131t
+
G\u00fcven d\u00fczeyi
+
Gerekirse ek tan\u0131lama sorusu<\/strong><\/p>\n\n\n\n

Dolay\u0131s\u0131yla sistem kesin bir etiket \u00fcretmek yerine bir tan\u0131 hipotezi<\/strong> olu\u015fturur.<\/p>\n\n\n\n

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

Olas\u0131 darbo\u011faz:<\/strong> PBIM-06
G\u00fcven:<\/strong> %68
Kan\u0131t:<\/strong> \u00d6\u011frencinin yan\u0131t\u0131ndaki belirli bir ifade
Gerekli ek kan\u0131t:<\/strong> Bir tan\u0131lama sorusuna verilecek yeni yan\u0131t<\/p>\n\n\n\n

Bu yakla\u015f\u0131m PBIM-AOS’u basit bir \u00f6\u011frenci s\u0131n\u0131fland\u0131rma sisteminden ay\u0131rmaktad\u0131r.<\/p>\n\n\n\n

Ve \u00e7ok \u00f6nemli ba\u015fka bir ilkeye g\u00f6t\u00fcrmektedir.<\/p>\n\n\n\n

Teacher-in-the-loop<\/h1>\n\n\n\n

PBIM-AOS’un amac\u0131 \u00f6\u011fretmenin pedagojik kararlar\u0131n\u0131 yapay zek\u00e2ya devretmek de\u011fildir.<\/p>\n\n\n\n

Tam tersine sistemin \u00f6nerdi\u011fi model \u015f\u00f6yledir:<\/p>\n\n\n\n

AI
\u2193
Tan\u0131 hipotezi
\u2193
Kan\u0131t
\u2193
Gerek\u00e7e
\u2193
\u00d6\u011fretmen
\u2199 \u2193 \u2198
Onayla \u2014 De\u011fi\u015ftir \u2014 Reddet
\u2193
M\u00fcdahale<\/strong><\/p>\n\n\n\n

Bu nedenle PBIM-AOS’u \u201cotomatik \u00f6\u011fretmen\u201d<\/strong> olarak de\u011fil, teacher-in-the-loop pedagogical decision support system<\/strong> olarak tan\u0131mlamak daha do\u011fru olacakt\u0131r.<\/p>\n\n\n\n

Yapay zek\u00e2 analiz eder.<\/p>\n\n\n\n

Yapay zek\u00e2 kan\u0131tlar\u0131 toplar.<\/p>\n\n\n\n

Yapay zek\u00e2 alternatifler \u00f6nerir.<\/p>\n\n\n\n

Ama kritik pedagojik karar\u0131n sorumlulu\u011fu \u00f6\u011fretmende kal\u0131r.<\/p>\n\n\n\n

\n\n\n\n

\u00d6\u011fretmenin rol\u00fc nas\u0131l de\u011fi\u015febilir?<\/h1>\n\n\n\n

Bu yakla\u015f\u0131m \u00f6\u011fretmenin \u00f6nemini azaltmak yerine rol\u00fcn\u00fc de\u011fi\u015ftirebilir.<\/p>\n\n\n\n

Geleneksel durumda \u00f6\u011fretmen \u00f6\u011frencinin problemini belirlemek, olas\u0131 nedenlerini d\u00fc\u015f\u00fcnmek, uygun etkinli\u011fi haz\u0131rlamak, \u00f6\u011frenciyi yeniden \u00f6l\u00e7mek ve sonu\u00e7lar\u0131 de\u011ferlendirmek zorundad\u0131r.<\/p>\n\n\n\n

PBIM-AOS destekli bir ortamda sistem bu s\u00fcrecin baz\u0131 b\u00f6l\u00fcmlerini haz\u0131rlayabilir:<\/p>\n\n\n\n

olas\u0131 \u00f6\u011frenme g\u00fc\u00e7l\u00fcklerini analiz etmek,
kan\u0131tlar\u0131 toplamak,
m\u00fcdahale se\u00e7enekleri olu\u015fturmak,
etkinlikler haz\u0131rlamak,
yeniden de\u011ferlendirme sorular\u0131 \u00fcretmek.<\/strong><\/p>\n\n\n\n

\u00d6\u011fretmen ise bu \u00e7\u0131kt\u0131lar\u0131 pedagojik bilgi ve deneyimiyle de\u011ferlendirerek nihai karar\u0131 verir. Kaynak belge bunu \u00f6\u011fretmenin karar verici rol\u00fcn\u00fc korurken rutin analiz ve haz\u0131rl\u0131k y\u00fck\u00fcn\u00fcn azalt\u0131lmas\u0131 olarak tan\u0131mlamaktad\u0131r.<\/p>\n\n\n\n

Bu anlamda \u00f6\u011fretmenin rol\u00fc i\u00e7erik \u00fcreten ki\u015fiden giderek daha fazla \u00f6\u011frenme s\u00fcrecini y\u00f6neten pedagojik karar vericiye<\/strong> d\u00f6n\u00fc\u015febilir.<\/p>\n\n\n\n

\n\n\n\n

Tek bir cevaptan \u00f6\u011frenme ge\u00e7mi\u015fine<\/h1>\n\n\n\n

PBIM-AOS’un bir ba\u015fka \u00f6nemli \u00f6zelli\u011fi, tek bir \u00f6\u011frenci yan\u0131t\u0131yla s\u0131n\u0131rl\u0131 kalmamas\u0131d\u0131r.<\/p>\n\n\n\n

PBIM a\u011f\u0131rl\u0131kl\u0131 olarak:<\/p>\n\n\n\n

Learning Outcome \u2192 Bottleneck<\/strong><\/p>\n\n\n\n

ili\u015fkisini yap\u0131land\u0131r\u0131rken, PBIM-AOS zaman i\u00e7erisinde:<\/p>\n\n\n\n

Student
\u2193
Learning Outcome
\u2193
Bottleneck
\u2193
Intervention
\u2193
Progress
\u2193
Bottleneck History<\/strong><\/p>\n\n\n\n

ili\u015fkisini kurabilir.<\/p>\n\n\n\n

Burada dikkat edilmesi gereken \u00f6nemli bir nokta vard\u0131r.<\/p>\n\n\n\n

Ama\u00e7 \u00f6\u011frenciye kal\u0131c\u0131 bir etiket vermek de\u011fildir.<\/p>\n\n\n\n

Bir \u00f6\u011frencinin \u201cPBIM-06 \u00f6\u011frencisi\u201d oldu\u011funu s\u00f6ylemek yerine, sistem zaman i\u00e7inde g\u00f6zlenen kan\u0131tlara dayal\u0131, de\u011fi\u015febilen bir learner model<\/strong> olu\u015fturmal\u0131d\u0131r.<\/p>\n\n\n\n

\u00c7\u00fcnk\u00fc \u00f6\u011frenme de \u00f6\u011frencinin ya\u015fad\u0131\u011f\u0131 g\u00fc\u00e7l\u00fckler de dinamiktir.<\/p>\n\n\n\n

\n\n\n\n

\u00dc\u00e7 katmanl\u0131 PBIM ekosistemi<\/h1>\n\n\n\n

Bu geli\u015fim PBIM ekosistemini \u00fc\u00e7 katmanda d\u00fc\u015f\u00fcnmemize olanak verir:<\/p>\n\n\n\n

Level 1 \u2014 PBIM Model<\/h3>\n\n\n\n

Perceptual Bottleneck Inventory for Mathematics<\/strong><\/p>\n\n\n\n

Bilimsel ve pedagojik \u00e7ekirdektir.<\/p>\n\n\n\n

\u00d6\u011frenme \u00e7\u0131kt\u0131lar\u0131, alg\u0131sal darbo\u011fazlar, tipik hatalar, tan\u0131lama ve m\u00fcdahale bilgilerini yap\u0131land\u0131r\u0131r.<\/p>\n\n\n\n

Level 2 \u2014 PBIM Platform<\/h3>\n\n\n\n

PBIM modelini \u00f6\u011fretmenler ve ara\u015ft\u0131rmac\u0131lar i\u00e7in kullan\u0131labilir hale getiren yaz\u0131l\u0131m katman\u0131d\u0131r.<\/p>\n\n\n\n

Level 3 \u2014 PBIM-AOS<\/h3>\n\n\n\n

PBIM Agentic Operating System<\/strong><\/p>\n\n\n\n

PBIM bilgisini kullanarak tan\u0131lama, kan\u0131t toplama, m\u00fcdahale, etkinlik, de\u011ferlendirme ve yeniden tan\u0131lama s\u00fcre\u00e7lerini koordine eden agentic zek\u00e2 katman\u0131d\u0131r.<\/p>\n\n\n\n

Kaynak belgede de PBIM ekosistemi tam olarak bu \u00fc\u00e7 d\u00fczeyle kavramsalla\u015ft\u0131r\u0131lmaktad\u0131r: PBIM Model bilimsel\/pedagojik \u00e7ekirdek, PBIM Platform modeli kullan\u0131labilir hale getiren yaz\u0131l\u0131m ve PBIM-AOS \u00f6\u011frenme s\u00fcre\u00e7lerini y\u00f6neten agentic zek\u00e2 katman\u0131d\u0131r.<\/p>\n\n\n\n

Bunu daha k\u0131sa bi\u00e7imde \u015f\u00f6yle ifade edebiliriz:<\/p>\n\n\n\n

PBI \u2192 kuramsal \u00e7er\u00e7eve<\/strong><\/p>\n\n\n\n

PBIM \u2192 matematik e\u011fitimi i\u00e7in yap\u0131land\u0131r\u0131lm\u0131\u015f pedagojik model<\/strong><\/p>\n\n\n\n

PBIM Platform \u2192 modelin \u00e7al\u0131\u015fan uygulamas\u0131<\/strong><\/p>\n\n\n\n

PBIM-AOS \u2192 modeli harekete ge\u00e7iren agentic pedagojik karar sistemi<\/strong><\/p>\n\n\n\n

\n\n\n\n

Yapay zek\u00e2 \u00f6\u011fretmenin yerine mi ge\u00e7iyor?<\/h1>\n\n\n\n

PBIM-AOS a\u00e7\u0131s\u0131ndan cevap hay\u0131r<\/strong>.<\/p>\n\n\n\n

As\u0131l hedef yapay zek\u00e2n\u0131n \u00f6\u011fretmen yerine karar vermesi de\u011fil, \u00f6\u011fretmenin karar verebilmesi i\u00e7in daha sistematik kan\u0131t \u00fcretmesidir.<\/p>\n\n\n\n

Bu ayr\u0131m e\u011fitimde yapay zek\u00e2n\u0131n gelece\u011fi a\u00e7\u0131s\u0131ndan \u00f6nemlidir.<\/p>\n\n\n\n

Sorulmas\u0131 gereken soru yaln\u0131zca:<\/p>\n\n\n\n

\n

\u201cYapay zek\u00e2 \u00f6\u011frenciye ne \u00f6\u011fretebilir?\u201d<\/strong><\/p>\n<\/blockquote>\n\n\n\n

de\u011fil,<\/p>\n\n\n\n

\n

\u201cYapay zek\u00e2 \u00f6\u011fretmenin \u00f6\u011frencinin \u00f6\u011frenmesini daha iyi anlamas\u0131na nas\u0131l yard\u0131mc\u0131 olabilir?\u201d<\/strong><\/p>\n<\/blockquote>\n\n\n\n

olmal\u0131d\u0131r.<\/p>\n\n\n\n

PBIM-AOS bu ikinci sorudan hareket etmektedir.<\/p>\n\n\n\n

Bu nedenle sistemin ba\u015far\u0131s\u0131 yaln\u0131zca kulland\u0131\u011f\u0131 b\u00fcy\u00fck dil modelinin g\u00fcc\u00fcyle \u00f6l\u00e7\u00fclmemelidir.<\/p>\n\n\n\n

Ba\u015far\u0131;<\/p>\n\n\n\n

tan\u0131lar\u0131n ne kadar kan\u0131ta dayal\u0131 oldu\u011fu,
\u00f6nerilerin ne kadar a\u00e7\u0131klanabilir oldu\u011fu,
\u00f6\u011fretmenin ne \u00f6l\u00e7\u00fcde kontrol sahibi oldu\u011fu,
m\u00fcdahalelerin \u00f6\u011frenmeye ne \u00f6l\u00e7\u00fcde katk\u0131 sa\u011flad\u0131\u011f\u0131<\/strong><\/p>\n\n\n\n

ile de\u011ferlendirilmelidir.<\/p>\n\n\n\n

\n\n\n\n

Sonu\u00e7: Cevap \u00fcreten AI’dan \u00f6\u011frenmeyi anlamaya \u00e7al\u0131\u015fan AI’a<\/h1>\n\n\n\n

\u00dcretken yapay zek\u00e2n\u0131n e\u011fitimdeki ilk d\u00f6nemi b\u00fcy\u00fck \u00f6l\u00e7\u00fcde cevap \u00fcretme<\/strong> \u00fczerine kuruluydu.<\/p>\n\n\n\n

Agentic AI ise farkl\u0131 bir olas\u0131l\u0131\u011f\u0131n kap\u0131s\u0131n\u0131 a\u00e7maktad\u0131r:<\/p>\n\n\n\n

\u00f6\u011frenme s\u00fcrecini izleyen, kan\u0131t toplayan, hipotez geli\u015ftiren, m\u00fcdahale se\u00e7enekleri olu\u015fturan ve sonu\u00e7lar\u0131 yeniden de\u011ferlendiren sistemler.<\/strong><\/p>\n\n\n\n

PBIM-AOS’un hedefledi\u011fi d\u00f6n\u00fc\u015f\u00fcm de budur.<\/p>\n\n\n\n

PBIM, \u00f6\u011frenci hatalar\u0131 ile bunlar\u0131n alt\u0131nda bulunabilecek alg\u0131sal darbo\u011fazlar aras\u0131ndaki pedagojik bilgiyi yap\u0131land\u0131rmaktad\u0131r.<\/p>\n\n\n\n

PBIM-AOS ise bu bilgiyi dinamik bir \u00f6\u011frenme s\u00fcrecinin par\u00e7as\u0131 haline getirmeyi ama\u00e7lamaktad\u0131r.<\/p>\n\n\n\n

B\u00f6ylece:<\/p>\n\n\n\n

\u00d6\u011frenci hatas\u0131 \u2192 Kan\u0131t \u2192 Alg\u0131sal darbo\u011faz hipotezi \u2192 \u00d6\u011fretmen karar\u0131 \u2192 M\u00fcdahale \u2192 Etkinlik \u2192 Yeni kan\u0131t \u2192 Yeniden de\u011ferlendirme<\/strong><\/p>\n\n\n\n

\u015feklinde s\u00fcrekli bir \u00f6\u011frenme d\u00f6ng\u00fcs\u00fc kurulabilir.<\/p>\n\n\n\n

Belki de e\u011fitimde agentic yapay zek\u00e2n\u0131n en \u00f6nemli potansiyeli burada yatmaktad\u0131r:<\/p>\n\n\n\n

\n

Yapay zek\u00e2n\u0131n \u00f6\u011fretmen yerine karar vermesi de\u011fil, \u00f6\u011fretmenin \u00f6\u011frencinin nas\u0131l d\u00fc\u015f\u00fcnd\u00fc\u011f\u00fcn\u00fc daha iyi g\u00f6rebilmesine yard\u0131mc\u0131 olmas\u0131.<\/strong><\/p>\n<\/blockquote>\n\n\n\n

\n\n\n\n

Prof. Dr. M. Ya\u015far \u00d6zden<\/strong>
Educational Technology & AI in Education<\/em>
Founder, PBI \u2014 Perceptual Bottleneck Inventory<\/em>
Developer, PBIM \u2014 Perceptual Bottleneck Inventory for Mathematics<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

PBIM-AOS, \u00f6\u011frenci hatalar\u0131n\u0131 yaln\u0131zca s\u0131n\u0131fland\u0131rmak yerine, bu hatalar\u0131n arkas\u0131ndaki alg\u0131sal darbo\u011fazlar\u0131 kan\u0131ta dayal\u0131 olarak inceleyen \u00e7ok ajanl\u0131 bir yapay zek\u00e2 yakla\u015f\u0131m\u0131d\u0131r. Bu yaz\u0131da PBIM\u2019in agentic AI mimarisine nas\u0131l d\u00f6n\u00fc\u015ft\u00fcr\u00fclebilece\u011fini ve \u00f6\u011fretmenin nihai karar verici olarak nas\u0131l sistemin merkezinde tutulabilece\u011fini ele al\u0131yorum.<\/p>\n","protected":false},"author":1,"featured_media":686,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3,4],"tags":[],"class_list":["post-684","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-genel-duyurular","category-guncel"],"acf":[],"_links":{"self":[{"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/posts\/684","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/comments?post=684"}],"version-history":[{"count":2,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/posts\/684\/revisions"}],"predecessor-version":[{"id":688,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/posts\/684\/revisions\/688"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/media\/686"}],"wp:attachment":[{"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/media?parent=684"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/categories?post=684"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/myozden.info\/index.php\/wp-json\/wp\/v2\/tags?post=684"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}