From mediawiki.org
This page is a translated version of the page Wikimedia Discovery/Knowledge Engine FAQ and the translation is 49% complete.


留意点としてコミュニティと前ウィキメディア財団事務長はこの主題に関して議論を交わしています:Knowledge Engine/FAQ(よくある質問)


「知のエンジン」("Knowledge Engine")とはかつての用語でウィキメディアのプロジェクト群を横断するコンテンツの検索と発見に関連した複数のイニシアティブを説明するため使われました。ナイト助成財団(Knight Foundation)がこれを参照し(「助成対象 > ジャーナリズム > ウィキペディアの知のエンジン」"What we fund / Journalism / Knowledge Engine By Wikipedia")次のように述べています。「ウィキペディアにおける知のエンジン開発の初期段階を支援し情報検索の新しいモデルを進展させるもので、(訳注:対象は)信頼でき依拠するに足りる公開の情報をインターネット上で発見するシステムである」"To advance new models for finding information by supporting stage one development of the Knowledge Engine by Wikipedia, a system for discovering reliable and trustworthy public information on the Internet."

助成金の詳細を開示するように要請があり、 Lila Tretikov から個人の見解と助成事業の内容と成果を共有しました。ナイト助成基金(The Knight Foundation)より2016年2月付で助成申請書の公開に承諾を得ています。

コンセプトが進化するにつれて、さまざまな発想が含まれ多くは不採用となりました。助成そのものは当初「ウィキペディアの進める知のエンジン」("Knowledge Engine by Wikipedia")を「インターネットの透明な検索エンジン第1号」("the Internet's first transparent search engine")と定義し、用語として「知のエンジン」("Knowledge Engine")も新企画の説明としてではなくそれを用いて発見チーム(Discovery team)の目指すところの説明に充てています。ウィキメディア財団は混乱が生じたとして、2015-16年次予算第三半期以降はこの用語の使用をやめています。


Google に対抗する検索エンジンを作成しているのか?

No, as we said when this FAQ was created in early November 2015, we are not creating a general internet search engine. The team has been tasked with improving a search function for Wikimedia sites.

In the sense that every site on the Internet competes for attention, and especially as a "go-to" resource for users, Wikipedia and its sister sites compete with Google and other search engines for certain kinds of inquiries. Namely, people who want to learn. Our visitors are people who want information and knowledge, and our mission is to serve them well. Our current search function does not serve our users well, so we are improving the existing CirrusSearch infrastructure with better relevance, multi language, multi project search and incorporating new data sources (like Maps) for our projects. We want a relevant and consistent experience for users across searches for both wikipedia.org and our project sites. Looking farther forward, we will explore including other sources of open knowledge. We remain fully committed to the movement's vision and values.

In the past some within the WMF talked about building a noncommercial search engine. The WMF thought about that direction but that did not go anywhere as an actual deliverable. By the time this page was created the focus was on the search function of Wikimedia projects and the joint announcement with Knight noted that the focus of initial research was improving the in-house function. That announcement also leaves open the possibility of broadening the reach of the program as it develops to include other publicly available sources of information.[1] In February 2016 the management of the WMF clarified that the WMF is "not building a global crawler search engine."

Nobody at the WMF intended for the search function at WMF sites to do things to help you find things to buy, as Google does, and there is no evidence that the WMF wanted to do all the many, many things that Google/Alphabet does.

Could search results in Wikipedia include more information from its sister projects?

The Wikimedia movement's vision is to make the sum of all human knowledge freely available to everyone. Wikipedia is our largest and most well-known project, but there are many other projects like Wikimedia Commons and Wikidata that move us towards making our vision happen. These projects have millions of users every month! So, can we make a search system that's good and meets the needs of our users and show content from around our movement? We think people would use it. As a result it would bring more attention to the great work in projects across the movement by using one of the larger, more well-known places people visit.

What licenses will those new data sources be under?

This will need more discussion as we want to be able to conform to the standards and policies of the Wikimedia projects they would need to serve. Our first exploration was with OSM licensing and legal and we'll want to learn from that in any further work.

Does that mean we are looking to shift search traffic away from third parties?

We love all the third party traffic that we get and we hope that it increases over time.

What we want is to improve people's ability to find what they want, once they are here. Too many times, users experience this:

  1. Search on Google, Bing, etc
  2. Follow Wikipedia Link
  3. Read
  4. Leave and search Google, Bing, etc again because you are specifically looking for more information but couldn't find it using the on-wiki search (CirrusSearch).

We want people who come to the Wikipedia and Wikimedia projects to find the information they want, and the more effective we are at serving them, the more they will come and the more they will stay. This means we are better serving our mission.

In addition, the WMF survives on donations, and the more people come directly to us, and the more they find what they need here, the more likely they are to donate.



ナイト助成基金(The Knight Foundation)はウィキメディアのプロジェクト群における検索の実態計測と改善に関して、ウィキメディア財団に予備調査(exploratory)の助成をつけました。


This grant does not increase the team's budget for this fiscal year.


  • 助成事業の募集 - ナイト助成基金のウェブサイト、掲載日不詳、支給期間の定義は2015年9月1日から2016年8月31日

Who is the Knight Foundation?

The Knight Foundation is a philanthropic organization dedicated to supporting media, journalism and fostering communities and the arts.

努力の一環として以前にもウィキメディア財団に対しウィキペディア・ゼロ(Wikipedia Zero)を支える助成金を発給しています。


  • 鍵となる質問に答えること:
    • もし単なる百科事典ではなくオープンなチャネルだった場合、利用者はウィキペディアを読みに来るか?
    • ウィキメディア財団はキャリアならびに相手先(委託者)ブランド名製造者(OEM=Original Equipment Manufacturers)経由でウィキペディアを(訳注:市場の製品に)組み込めるか?
  • 重要業績評価指標(KPI=Key Performance Indicators)を用いて製品の反復開発に通知し、プロトタイプの理解と機能開発に関する鍵となる理解を確立
  • 利用者登録をしそうな対象に試験を行う
  • 鍵となるKPIについて、一般社会に向いたダッシュボードを作成
  • 計測対象:
    • 利用者の満足度(どのクエリで関連するコンテンツが浮上したか分析する)
    • 利用者側で認知したロードタイム
    • 結果がヒットしなかったレート
    • API(アプリケーション プログラミング インターフェイス)の利用


  • 浮上したコンテンツの関連(性)すなわちレレバンシーを調べた試験結果
  • 調査ならびに利用者テストの試験結果
  • ウィキペディアの検索に用いる検索エンジンならびに API の改良
  • 一般社会に向いたダッシュボードとして製品開発におけるコア計測指標を提示
  • 可能性の例示となる小規模なデータセットを用いたプロトタイプ見本
  • 検索と発見の体験について現状の利用者の理解のフローを対象にした利用者テストと調査


  1. The release says: "Over the last decade, the world has seen a surge in digital information. People can access large amounts of information online, mostly through a small number of closed technologies. Through this project, the Wikimedia Foundation will test ways to make relevant information more accessible and investigate transparent methods for collecting, connecting and retrieving this information consistent with the values of Wikipedia and the open Web."