{"id":475934,"date":"2023-08-09T07:24:43","date_gmt":"2023-08-09T07:24:43","guid":{"rendered":""},"modified":"2023-09-05T11:11:39","modified_gmt":"2023-09-05T11:11:39","slug":"attention-mechanism","status":"publish","type":"wiki","link":"https:\/\/oneproxy.pro\/tr\/wiki\/attention-mechanism\/","title":{"rendered":"Dikkat mekanizmas\u0131"},"content":{"rendered":"

Dikkat mekanizmas\u0131, derin \u00f6\u011frenme ve yapay zeka alan\u0131nda \u00e7ok \u00f6nemli bir kavramd\u0131r. Bir modelin dikkatini girdi verilerinin belirli b\u00f6l\u00fcmlerine odaklamas\u0131na izin vererek, en alakal\u0131 bilgilere daha fazla kaynak ay\u0131rmas\u0131n\u0131 sa\u011flayarak \u00e7e\u015fitli g\u00f6revlerin performans\u0131n\u0131 art\u0131rmak i\u00e7in kullan\u0131lan bir mekanizmad\u0131r. Ba\u015flang\u0131\u00e7ta insan\u0131n bili\u015fsel s\u00fcre\u00e7lerinden ilham alan Dikkat mekanizmas\u0131, do\u011fal dil i\u015fleme, bilgisayarl\u0131 g\u00f6rme ve s\u0131ral\u0131 veya mekansal bilginin \u00e7ok \u00f6nemli oldu\u011fu di\u011fer alanlarda yayg\u0131n uygulamalar bulmu\u015ftur.<\/p>\n

Dikkat Mekanizmas\u0131n\u0131n K\u00f6keni ve \u0130lk S\u00f6z\u00fc<\/h2>\n

Dikkat fikrinin k\u00f6keni psikoloji alan\u0131nda 20. y\u00fczy\u0131l\u0131n ba\u015flar\u0131na kadar uzanmaktad\u0131r. Psikologlar William James ve John Dewey se\u00e7ici dikkat ve bilin\u00e7 kavramlar\u0131n\u0131 ara\u015ft\u0131rarak Dikkat mekanizmas\u0131n\u0131n nihai geli\u015fiminin temelini att\u0131lar.<\/p>\n

Derin \u00f6\u011frenme ba\u011flam\u0131nda Dikkat mekanizmas\u0131ndan ilk s\u00f6z Bahdanau ve arkada\u015flar\u0131n\u0131n \u00e7al\u0131\u015fmas\u0131na atfedilebilir. (2014), \u201cDikkat Temelli Sinir Makinesi \u00c7evirisi\u201d modelini tan\u0131tan ki\u015fidir. Bu, makine \u00e7evirisinde \u00f6nemli bir ilerlemeye i\u015faret ediyordu; model, \u00e7\u0131kt\u0131 c\u00fcmlesinde kar\u015f\u0131l\u0131k gelen kelimeleri \u00fcretirken, girdi c\u00fcmlesindeki belirli kelimelere se\u00e7ici olarak odaklanabiliyordu.<\/p>\n

Dikkat Mekanizmas\u0131 Hakk\u0131nda Detayl\u0131 Bilgi: Konuyu Geni\u015fletmek<\/h2>\n

Dikkat mekanizmas\u0131n\u0131n birincil hedefi, t\u00fcm girdi verilerinin sabit uzunluklu bir temsile kodlanmas\u0131 y\u00fck\u00fcn\u00fc azaltarak derin \u00f6\u011frenme modellerinin verimlili\u011fini ve etkinli\u011fini art\u0131rmakt\u0131r. Bunun yerine, eldeki g\u00f6rev i\u00e7in gerekli olan girdi verilerinin en ilgili k\u0131s\u0131mlar\u0131na odaklanmaya odaklan\u0131r. Bu \u015fekilde model \u00f6nemli bilgilere odaklanabilir, daha do\u011fru tahminler yapabilir ve daha uzun dizileri verimli bir \u015fekilde i\u015fleyebilir.<\/p>\n

Dikkat mekanizmas\u0131n\u0131n ard\u0131ndaki temel fikir, girdi ve \u00e7\u0131kt\u0131 dizilerinin \u00f6\u011feleri aras\u0131nda yumu\u015fak bir hizalama sa\u011flamakt\u0131r. Girdi dizisinin her bir \u00f6\u011fesine farkl\u0131 \u00f6nem a\u011f\u0131rl\u0131klar\u0131 atar ve her bir \u00f6\u011fenin, modelin \u00e7\u0131kt\u0131 \u00fcretiminin ge\u00e7erli ad\u0131m\u0131yla olan ili\u015fkisini yakalar.<\/p>\n

Dikkat Mekanizmas\u0131n\u0131n \u0130\u00e7 Yap\u0131s\u0131: Nas\u0131l \u00c7al\u0131\u015f\u0131r?<\/h2>\n

Dikkat mekanizmas\u0131 tipik olarak \u00fc\u00e7 ana bile\u015fenden olu\u015fur:<\/p>\n

    \n
  1. \n

    Sorgu<\/strong>: Bu, \u00e7\u0131k\u0131\u015f s\u0131ras\u0131ndaki ge\u00e7erli ad\u0131m\u0131 veya konumu temsil eder.<\/p>\n<\/li>\n

  2. \n

    Anahtar<\/strong>: Bunlar modelin ilgilenece\u011fi girdi dizisinin unsurlar\u0131d\u0131r.<\/p>\n<\/li>\n

  3. \n

    De\u011fer<\/strong>: Bunlar, ba\u011flam vekt\u00f6r\u00fcn\u00fc hesaplamak i\u00e7in kullan\u0131lan bilgileri sa\u011flayan, her bir anahtarla ili\u015fkili kar\u015f\u0131l\u0131k gelen de\u011ferlerdir.<\/p>\n<\/li>\n<\/ol>\n

    Dikkat s\u00fcreci, sorgu ile t\u00fcm anahtarlar aras\u0131ndaki ilgi veya dikkat a\u011f\u0131rl\u0131klar\u0131n\u0131n hesaplanmas\u0131n\u0131 i\u00e7erir. Bu a\u011f\u0131rl\u0131klar daha sonra ba\u011flam vekt\u00f6r\u00fcn\u00fc olu\u015fturarak de\u011ferlerin a\u011f\u0131rl\u0131kl\u0131 toplam\u0131n\u0131 hesaplamak i\u00e7in kullan\u0131l\u0131r. Bu ba\u011flam vekt\u00f6r\u00fc, ge\u00e7erli ad\u0131mda son \u00e7\u0131kt\u0131y\u0131 \u00fcretmek i\u00e7in sorguyla birle\u015ftirilir.<\/p>\n

    Dikkat Mekanizmas\u0131n\u0131n Temel \u00d6zelliklerinin Analizi<\/h2>\n

    Dikkat mekanizmas\u0131, yayg\u0131n olarak benimsenmesine katk\u0131da bulunan \u00e7e\u015fitli temel \u00f6zellikler ve avantajlar sunmaktad\u0131r:<\/p>\n

      \n
    1. \n

      Esneklik<\/strong>: Dikkat uyarlanabilir ve makine \u00e7evirisi, duygu analizi, g\u00f6r\u00fcnt\u00fc altyaz\u0131s\u0131 ve konu\u015fma tan\u0131ma dahil olmak \u00fczere \u00e7e\u015fitli derin \u00f6\u011frenme g\u00f6revlerine uygulanabilir.<\/p>\n<\/li>\n

    2. \n

      Paralellik<\/strong>: Geleneksel s\u0131ral\u0131 modellerin aksine, Dikkat tabanl\u0131 modeller giri\u015f verilerini paralel olarak i\u015fleyerek e\u011fitim s\u00fcresini \u00f6nemli \u00f6l\u00e7\u00fcde azalt\u0131r.<\/p>\n<\/li>\n

    3. \n

      Uzun menzilli ba\u011f\u0131ml\u0131l\u0131klar<\/strong>: Dikkat, s\u0131ral\u0131 verilerdeki uzun vadeli ba\u011f\u0131ml\u0131l\u0131klar\u0131n yakalanmas\u0131na yard\u0131mc\u0131 olarak daha iyi anla\u015f\u0131lmas\u0131na ve ilgili \u00e7\u0131kt\u0131lar\u0131n olu\u015fturulmas\u0131na olanak tan\u0131r.<\/p>\n<\/li>\n

    4. \n

      Yorumlanabilirlik<\/strong>: Dikkat mekanizmalar\u0131, modelin girdi verilerinin hangi b\u00f6l\u00fcmlerini en alakal\u0131 olarak de\u011ferlendirdi\u011fine dair i\u00e7g\u00f6r\u00fc sa\u011flayarak yorumlanabilirli\u011fi art\u0131r\u0131r.<\/p>\n<\/li>\n<\/ol>\n

      Dikkat Mekanizmas\u0131 T\u00fcrleri<\/h2>\n

      Her biri belirli g\u00f6revlere ve veri yap\u0131lar\u0131na g\u00f6re uyarlanm\u0131\u015f farkl\u0131 t\u00fcrde Dikkat mekanizmalar\u0131 vard\u0131r. Yayg\u0131n t\u00fcrlerden baz\u0131lar\u0131 \u015funlard\u0131r:<\/p>\n\n\n\n\n\n\n\n\n
      Tip<\/th>\nTan\u0131m<\/th>\n<\/tr>\n<\/thead>\n
      K\u00fcresel Dikkat<\/strong><\/td>\nDikkat \u00e7ekmek i\u00e7in giri\u015f s\u0131ras\u0131n\u0131n t\u00fcm \u00f6\u011felerini dikkate al\u0131r.<\/td>\n<\/tr>\n
      Yerel \u0130lgi<\/strong><\/td>\nGiri\u015f s\u0131ras\u0131ndaki yaln\u0131zca s\u0131n\u0131rl\u0131 say\u0131da \u00f6\u011feye odaklan\u0131r.<\/td>\n<\/tr>\n
      Ki\u015fisel Dikkat<\/strong><\/td>\nTransformat\u00f6r mimarilerinde yayg\u0131n olarak kullan\u0131lan, ayn\u0131 s\u0131ra i\u00e7erisinde farkl\u0131 konumlara kat\u0131l\u0131r.<\/td>\n<\/tr>\n
      \u00d6l\u00e7eklendirilmi\u015f Nokta \u00dcr\u00fcn Dikkati<\/strong><\/td>\nKaybolan\/patlayan degradeleri \u00f6nlemek i\u00e7in \u00f6l\u00e7eklenen dikkat a\u011f\u0131rl\u0131klar\u0131n\u0131 hesaplamak i\u00e7in nokta \u00e7arp\u0131m\u0131 kullan\u0131r.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      Dikkat Mekanizmas\u0131n\u0131 Kullanma Yollar\u0131, Sorunlar ve \u00c7\u00f6z\u00fcmler<\/h2>\n

      Dikkat mekanizmas\u0131n\u0131n \u00e7e\u015fitli uygulamalar\u0131 vard\u0131r; bunlardan baz\u0131lar\u0131 \u015funlard\u0131r:<\/p>\n

        \n
      1. \n

        Makine \u00c7evirisi<\/strong>: Dikkate dayal\u0131 modeller, \u00e7eviri s\u0131ras\u0131nda ilgili kelimelere odaklanarak makine \u00e7evirisini \u00f6nemli \u00f6l\u00e7\u00fcde geli\u015ftirmi\u015ftir.<\/p>\n<\/li>\n

      2. \n

        Resim Altyaz\u0131s\u0131<\/strong>: Bilgisayarla g\u00f6rme g\u00f6revlerinde Dikkat, g\u00f6r\u00fcnt\u00fcn\u00fcn farkl\u0131 b\u00f6l\u00fcmlerine se\u00e7ici olarak m\u00fcdahale ederek a\u00e7\u0131klay\u0131c\u0131 altyaz\u0131lar olu\u015fturulmas\u0131na yard\u0131mc\u0131 olur.<\/p>\n<\/li>\n

      3. \n

        Konu\u015fma tan\u0131ma<\/strong>: Dikkat, akustik sinyalin \u00f6nemli k\u0131s\u0131mlar\u0131na odaklanarak konu\u015fman\u0131n daha iyi tan\u0131nmas\u0131n\u0131 sa\u011flar.<\/p>\n<\/li>\n<\/ol>\n

        Ancak Dikkat mekanizmalar\u0131 a\u015fa\u011f\u0131daki gibi zorluklarla da kar\u015f\u0131 kar\u015f\u0131yad\u0131r:<\/p>\n

          \n
        1. \n

          Hesaplamal\u0131 Karma\u015f\u0131kl\u0131k<\/strong>: Uzun bir s\u0131radaki t\u00fcm \u00f6\u011felerle ilgilenmek hesaplama a\u00e7\u0131s\u0131ndan pahal\u0131 olabilir.<\/p>\n<\/li>\n

        2. \n

          A\u015f\u0131r\u0131 uyum g\u00f6sterme<\/strong>: Dikkat bazen verilerdeki g\u00fcr\u00fclt\u00fcy\u00fc ezberleyebilir ve bu da a\u015f\u0131r\u0131 uyumla sonu\u00e7lanabilir.<\/p>\n<\/li>\n<\/ol>\n

          Bu sorunlar\u0131n \u00e7\u00f6z\u00fcmleri a\u015fa\u011f\u0131daki gibi tekniklerin kullan\u0131lmas\u0131n\u0131 i\u00e7erir: seyrekli\u011fe neden olan dikkat<\/strong>, \u00e7ok kafal\u0131 dikkat<\/strong> farkl\u0131 desenleri yakalamak ve d\u00fczenlile\u015ftirme<\/strong> a\u015f\u0131r\u0131 uyumu \u00f6nlemek i\u00e7in.<\/p>\n

          Ana \u00d6zellikler ve Benzer Terimlerle Kar\u015f\u0131la\u015ft\u0131rmalar<\/h2>\n\n\n\n\n\n\n\n\n
          karakteristik<\/th>\nDikkat Mekanizmas\u0131<\/th>\nBenzer Terimler (\u00f6rn. Odaklanma, Se\u00e7meli \u0130\u015fleme)<\/th>\n<\/tr>\n<\/thead>\n
          Ama\u00e7<\/strong><\/td>\n\u0130lgili bilgilere odaklanarak model performans\u0131n\u0131 iyile\u015ftirin.<\/td>\nBenzer ama\u00e7 ancak sinir a\u011f\u0131 entegrasyonundan yoksun olabilir.<\/td>\n<\/tr>\n
          Bile\u015fenler<\/strong><\/td>\nSorgu, Anahtar, De\u011fer<\/td>\nBenzer bile\u015fenler mevcut olabilir ancak mutlaka ayn\u0131 olmas\u0131 gerekmez.<\/td>\n<\/tr>\n
          Uygulamalar<\/strong><\/td>\nNLP, Bilgisayarl\u0131 G\u00f6rme, Konu\u015fma Tan\u0131ma vb.<\/td>\nBenzer uygulamalar, ancak baz\u0131 durumlarda o kadar etkili de\u011fil.<\/td>\n<\/tr>\n
          Yorumlanabilirlik<\/strong><\/td>\n\u0130lgili giri\u015f verilerine ili\u015fkin \u00f6ng\u00f6r\u00fcler sa\u011flar.<\/td>\nYorumlanabilirlik d\u00fczeyi benzerdir ancak dikkat daha a\u00e7\u0131kt\u0131r.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

          Dikkat Mekanizmas\u0131na \u0130li\u015fkin Perspektifler ve Gelecek Teknolojiler<\/h2>\n

          Dikkat mekanizmas\u0131 geli\u015fmeye devam ediyor ve Dikkat ile ilgili gelecekteki teknolojiler \u015funlar\u0131 i\u00e7erebilir:<\/p>\n

            \n
          1. \n

            Seyrek Dikkat<\/strong>: Girdideki yaln\u0131zca ilgili \u00f6\u011felere odaklanarak hesaplama verimlili\u011fini art\u0131rma teknikleri.<\/p>\n<\/li>\n

          2. \n

            Hibrit Modeller<\/strong>: Geli\u015fmi\u015f performans i\u00e7in Dikkatin haf\u0131za a\u011flar\u0131 veya takviyeli \u00f6\u011frenme gibi di\u011fer tekniklerle entegrasyonu.<\/p>\n<\/li>\n

          3. \n

            Ba\u011flamsal Dikkat<\/strong>: Davran\u0131\u015flar\u0131n\u0131 ba\u011flamsal bilgiye dayal\u0131 olarak uyarlanabilir bir \u015fekilde ayarlayan dikkat mekanizmalar\u0131.<\/p>\n<\/li>\n<\/ol>\n

            Proxy Sunucular Nas\u0131l Kullan\u0131labilir veya Dikkat Mekanizmas\u0131 ile \u0130li\u015fkilendirilebilir<\/h2>\n

            Proxy sunucular\u0131, \u00f6nbelle\u011fe alma, g\u00fcvenlik ve anonimlik gibi \u00e7e\u015fitli i\u015flevler sa\u011flayarak istemciler ve internet aras\u0131nda arac\u0131 g\u00f6revi g\u00f6r\u00fcr. Proxy sunucular\u0131 ile Dikkat mekanizmas\u0131 aras\u0131ndaki do\u011frudan ili\u015fki belirgin olmasa da, Dikkat mekanizmas\u0131 OneProxy (oneproxy.pro) gibi proxy sunucu sa\u011flay\u0131c\u0131lar\u0131na a\u015fa\u011f\u0131daki \u015fekillerde dolayl\u0131 olarak fayda sa\u011flayabilir:<\/p>\n

              \n
            1. \n

              Kaynak Tahsisi<\/strong>: Dikkat'i kullanarak, proxy sunucular kaynaklar\u0131 daha verimli bir \u015fekilde tahsis edebilir, en ilgili isteklere odaklanabilir ve sunucu performans\u0131n\u0131 optimize edebilir.<\/p>\n<\/li>\n

            2. \n

              Uyarlanabilir \u00d6nbelle\u011fe Alma<\/strong>: Proxy sunucular\u0131, s\u0131k istenen i\u00e7eri\u011fi belirlemek ve daha h\u0131zl\u0131 eri\u015fim i\u00e7in ak\u0131ll\u0131 bir \u015fekilde \u00f6nbelle\u011fe almak \u00fczere Dikkat'i kullanabilir.<\/p>\n<\/li>\n

            3. \n

              Anomali tespiti<\/strong>: Anormal isteklerin tespit edilmesi ve ele al\u0131nmas\u0131, proxy sunucular\u0131n\u0131n g\u00fcvenli\u011finin artt\u0131r\u0131lmas\u0131 konusunda dikkatli olunabilir.<\/p>\n<\/li>\n<\/ol>\n

              \u0130lgili Ba\u011flant\u0131lar<\/h2>\n

              Dikkat mekanizmas\u0131 hakk\u0131nda daha fazla bilgi i\u00e7in a\u015fa\u011f\u0131daki kaynaklara ba\u015fvurabilirsiniz:<\/p>\n

                \n
              1. Bahdanau ve di\u011ferleri, Ortakla\u015fa Hizalamay\u0131 ve \u00c7evirmeyi \u00d6\u011frenme yoluyla Sinir Makinesi \u00c7evirisi, 2014<\/a><\/li>\n
              2. Vaswani ve di\u011ferleri, \u0130htiyac\u0131n\u0131z Olan Tek \u015eey Dikkat, 2017<\/a><\/li>\n
              3. Chorowski ve di\u011ferleri, Konu\u015fma Tan\u0131ma i\u00e7in Dikkat Temelli Modeller, 2015<\/a><\/li>\n
              4. Xu ve di\u011ferleri, G\u00f6ster, Kat\u0131l ve Anlat: G\u00f6rsel Dikkat ile N\u00f6ral G\u00f6r\u00fcnt\u00fc Altyaz\u0131s\u0131 Olu\u015fturma, 2015<\/a><\/li>\n<\/ol>\n

                Sonu\u00e7 olarak, Dikkat mekanizmas\u0131 derin \u00f6\u011frenmede temel bir ilerlemeyi temsil ediyor ve modellerin ilgili bilgilere odaklanmas\u0131n\u0131 ve \u00e7e\u015fitli alanlarda performans\u0131 art\u0131rmas\u0131n\u0131 sa\u011fl\u0131yor. Makine \u00e7evirisi, g\u00f6rsel altyaz\u0131lama ve daha bir\u00e7ok alandaki uygulamalar\u0131 yapay zeka teknolojilerinde dikkate de\u011fer ilerlemelere yol a\u00e7t\u0131. Dikkat mekanizmas\u0131 alan\u0131 geli\u015fmeye devam ettik\u00e7e, OneProxy gibi proxy sunucu sa\u011flay\u0131c\u0131lar\u0131 kaynak tahsisini, \u00f6nbelle\u011fe almay\u0131 ve g\u00fcvenlik \u00f6nlemlerini geli\u015ftirmek ve kullan\u0131c\u0131lar\u0131na en iyi hizmeti sa\u011flamak i\u00e7in bu teknolojiden yararlanabilirler.<\/p>","protected":false},"featured_media":467660,"menu_order":0,"template":"","meta":{"_acf_changed":false,"content-type":"","inline_featured_image":false,"footnotes":""},"class_list":["post-475934","wiki","type-wiki","status-publish","has-post-thumbnail","hentry"],"acf":{"faq_title":"Frequently Asked Questions about Attention Mechanism: Enhancing Proxy Server Performance<\/mark>","faq_items":[{"question":"What is the Attention mechanism?","answer":"

                The Attention mechanism is a pivotal concept in deep learning and AI, allowing models to focus on the most relevant information in the input data. It enhances performance across various tasks, such as machine translation, image captioning, and speech recognition, by allocating resources more efficiently.<\/p>"},{"question":"How did the Attention mechanism originate?","answer":"

                The idea of attention can be traced back to early psychology studies on selective attention and consciousness by William James and John Dewey. In the context of deep learning, the Attention mechanism was first introduced in 2014 by Bahdanau et al. as part of a neural machine translation model.<\/p>"},{"question":"How does the Attention mechanism work?","answer":"

                The Attention mechanism involves three main components: Query, Key, and Value. It calculates relevance or attention weights between the Query and all Keys, then generates a context vector through a weighted sum of the Values. This context vector is combined with the Query to produce the final output.<\/p>"},{"question":"What are the key features of the Attention mechanism?","answer":"

                The Attention mechanism offers flexibility, parallelism, and the ability to capture long-range dependencies in data. It also provides interpretability, as it reveals which parts of the input data the model deems most important.<\/p>"},{"question":"What are the types of Attention mechanisms?","answer":"

                There are different types of Attention mechanisms, including Global Attention, Local Attention, Self-Attention, and Scaled Dot-Product Attention. Each type is suited for specific tasks and data structures.<\/p>"},{"question":"How can the Attention mechanism be used?","answer":"

                The Attention mechanism has various applications, including machine translation, image captioning, and speech recognition. It helps improve performance in these tasks by focusing on relevant information.<\/p>"},{"question":"What are the challenges of using the Attention mechanism?","answer":"

                Some challenges include computational complexity when attending to long sequences and the potential for overfitting. Solutions involve sparsity-inducing attention and regularization techniques.<\/p>"},{"question":"How does the Attention mechanism compare to similar terms?","answer":"

                The Attention mechanism is similar to the concept of focus or selective processing, but it stands out for its integration into neural network architectures and its explicit attention to relevant data.<\/p>"},{"question":"What are the future technologies related to the Attention mechanism?","answer":"

                Future technologies include sparse attention for improved efficiency, hybrid models integrating attention with other techniques, and contextual attention that adapts based on context.<\/p>"},{"question":"How can proxy servers benefit from the Attention mechanism?","answer":"

                Proxy servers like OneProxy can indirectly benefit from the Attention mechanism by optimizing resource allocation, adaptive caching, and improving anomaly detection for enhanced security.<\/p>"}]},"_links":{"self":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/wiki\/475934","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/wiki"}],"about":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/types\/wiki"}],"version-history":[{"count":0,"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/wiki\/475934\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/media\/467660"}],"wp:attachment":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/media?parent=475934"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}