Phabricator のプロジェクト名 #edit-review-improvements

ヘルプ:投稿の査読の改善/用語集

From MediaWiki.org
Jump to navigation Jump to search
This page is a translated version of the page Help:Edit Review Improvements/Glossary and the translation is 76% complete.

Other languages:
English • ‎français • ‎italiano • ‎polski • ‎português do Brasil • ‎čeština • ‎日本語

「投稿査読の改善」プロジェクト (ERI) に関連する用語を集めました。 斜体で記した用語はそれぞれの用語集にリンクします。

ERI のコンポーネントと特殊な用語

投稿の査読の改善
別名 ERI
編集の査読改善(Edit Review ImprovementsERI)とはCollaborationチームが進める改善プロジェクトで、現状の編集の査読プロセスが編集の初学者に与えかねない悪影響の低減と—編集の作動全体の底上げを同時に目指しています。
投稿の査読に使う新しいフィルタ
あるいは「特別ページ:最近の更新」の絞込み機能とも呼ばれます。
ERIの初期のリリース版では、特別ページである「最近の更新」の改善をいくつか試みており、査読者全般の対象選びと作業効率のサポートを目指しました。それらの改善は、中でも新人の寄稿者に役立つ可能性を備えています。
査読ストリーム
査読ストリーム(ReviewStream)は機械可読なフィードで、さまざまな編集の査読用ツールで利用するよう設計してあります。RCStreamと同様にMediaWikiのウィキ群の最近の更新を広めます。ReviewStreamはRCStreamの蓄えた情報にデータを追加し、編集から査読までの過程の改善を目指しています。
フィルター検索のツールバー
「最近の更新ページの改善」の一環。ユーザーはこのバーを押して「最近の更新」の絞りこみができ、「フィルターのドロップダウンパネル」(下図参照)から選ぶか、もしくは検索窓に入力します。
フィルターのドロップダウンパネル
「最近の更新ページの改善」の一環。絞りこみのオプションはフィルターパネルに表示され、ドロップダウン式のメニューから選びます。
使用中のフィルターの表示枠
「最近の更新ページの改善」の一環。使用中のフィルターを表示する枠を見ると、自分が使用中のフィルターや強調表示が一目で把握できます。
新人
A new user class defined for ERI denoting registered editors who are very new, having fewer than ten edits and four days of activity. Created because research shows that such editors are particularly vulnerable to rejections. Also the name of a filter in ERI projects.
初学者
A term defined for ERI denoting registered editors who have more days of activity and edits than “Newcomers” but fewer than “Experienced Users”. Also the name of a filter in ERI projects. On English Wikipedia, this category corresponds to Autoconfirmed.
利用歴の長いユーザー
A term defined for ERI denoting registered editors who have more than 30 days of activity and 500 edits. Also the name of a filter in ERI projects. On English Wikipedia, this category corresponds to Extended Confirmed.

ERI の定義とウィキにおける位置づけ

人工知能 (Artificial Intelligence)
ERI uses predictive filters powered by the artificial intelligence (AI) program ORES. Colloquially, the term artificial intelligence is applied when a machine mimics cognitive functions that humans associate with other human minds, such as natural language processing, problem solving or learning. AI is a broad term that’s closely associated with the more specific concept of “machine learning”.
機械学習 (Machine Learning)
Machine learning is a subfield of artificial intelligence that seeks to give computers the ability to learn tasks they haven’t been specifically programmed to do. ERI uses predictive filters powered by the machine-learning program ORES.
精度
ERI tools offer AI predictive filters at various levels of “accuracy”. Used in this context, accuracy refers to the percentage of search results the filter returns that correctly identify the property the filter is targeting. In other words, the more accurate the filter is, the fewer false positives it produces. (In the AI field, the correct technical term for this is “precision”.)
フィード (Feed)
ストリームとも呼ばれます。
A computer feed is a data format used for providing users with frequently updated content. In ERI, ReviewStream is a feed that broadcasts data about recent changes to anti-vandalism and edit-review programs.
善意に解釈する
ERI projects offer User Intent filters, which predict whether an edit was made in good faith. In this sense, the term refers to whether a user’s intention was to contribute to the wiki or damage it. A key goal of ERI is enabling reviewers to find and help editors — especially newcomers — who are trying to contribute in good faith but who may lack the skill or knowledge to do so successfully. (ERI’s User Intent filters are powered by the machine-learning program ORES, which includes a good-faith test.)
損害
ERI projects offer Contribution Quality filters, which predict whether an edit improves the wiki or damages it. In this sense, “damage” is a general term that covers both vandalism and unintentional errors of formatting, style, fact, etc. A key goal of ERI is enabling reviewers to find and help editors — especially newcomers — who are trying to contribute in good faith but whose lack of skill or knowledge leads them to cause damage. (ERI’s User Intent filters are powered by the machine-learning program ORES, which includes a “damaging” test.)