[{"data":1,"prerenderedAt":222},["ShallowReactive",2],{"topic-page:\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications\u002Frag-and-vector-databases":3},{"topic":4,"category":182,"seo":186,"breadcrumbs":191,"prerequisites":201,"nextTopics":202,"relatedTopics":210},{"id":5,"slug":6,"title":7,"summary":8,"canonicalPath":9,"categoryPath":10,"difficulty":11,"targetAudience":12,"estimatedMinutes":13,"generatedAt":14,"publishedAt":14,"updatedAt":14,"reviewStatus":15,"tags":16,"expectedKnowledge":26,"learningGoal":29,"keyTerms":30,"blocks":94},"it-aws-aws-certified-ai-practitioner-foundation-model-applications-rag-and-vector-databases","rag-and-vector-databases","RAGとベクトルデータベース","RAGとベクトルデータベースを、AIF-C01の公式objective 3.1.3・3.1.4に沿って、用語、判断順序、例、誤解修正で整理します。","\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications\u002Frag-and-vector-databases","\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications","beginner","cert_candidate",6,"2026-08-06","ai_generated",[17,18,19,20,21,22,23,24,25],"AWS Certified AI Practitioner","AIF-C01","基盤モデルの活用","RAG","埋め込み","ベクトルデータベース","Amazon Bedrock Knowledge Bases","OpenSearch","Aurora",[27,28],"AWSクラウドの基本用語を聞いたことがある","AIを使う業務例を日常的な言葉で考えられる","RAGとベクトルデータベースについて、試験で問われる違いと選択基準を説明し、短い業務例へ適用できる。",[31,39,43,49,52,56,59,66,70,74,78,82,86,90],{"id":32,"label":33,"shortDefinition":34,"supplement":35,"role":38},"retrieval-augmented-generation","検索拡張生成（RAG）","関連情報を検索し、その結果を文脈に加えて回答を生成する方式です。",{"kind":36,"text":37},"misconception","検索結果が正しいか、回答の根拠確認は必要です。","primary",{"id":40,"label":41,"shortDefinition":42,"role":38},"chunking","チャンク化","長いデータを、検索や処理に適した意味のまとまりへ分割することです。",{"id":44,"label":21,"shortDefinition":45,"supplement":46,"role":38},"embedding","文章や画像などの意味・特徴を数値ベクトルへ写した表現です。",{"kind":47,"text":48},"comparison","埋め込みは回答文ではなく、比較しやすい数値表現です。",{"id":50,"label":22,"shortDefinition":51,"role":38},"vector-database","埋め込みベクトルを保存し、近さに基づく検索を行うデータ基盤です。",{"id":53,"label":54,"shortDefinition":55,"role":38},"semantic-search","意味検索","語句の一致だけでなく、意味の近さを使って関連情報を探す検索です。",{"id":57,"label":23,"shortDefinition":58,"role":38},"amazon-bedrock-knowledge-bases","外部データの取得と基盤モデルによる生成を組み合わせるAWS機能です。",{"id":60,"label":61,"shortDefinition":62,"supplement":63,"role":65},"foundation-model","基盤モデル","大規模なデータで事前学習され、多様な下流タスクへ適応できるモデルです。",{"kind":36,"text":64},"すべての定型予測を、基盤モデルへ置き換える必要はありません。","contextual",{"id":67,"label":68,"shortDefinition":69,"role":65},"vector","ベクトル","複数の数値を順序付きで並べ、特徴や意味を表す表現です。",{"id":71,"label":72,"shortDefinition":73,"role":65},"amazon-opensearch-service","Amazon OpenSearch Service","検索・分析とベクトル検索を提供するAWSのマネージドサービスです。",{"id":75,"label":76,"shortDefinition":77,"role":65},"amazon-aurora","Amazon Aurora","MySQLおよびPostgreSQL互換のAWSマネージドリレーショナルデータベースです。",{"id":79,"label":80,"shortDefinition":81,"role":65},"amazon-neptune","Amazon Neptune","グラフデータの保存・検索を支えるAWSのマネージドデータベースです。",{"id":83,"label":84,"shortDefinition":85,"role":65},"amazon-rds-for-postgresql","Amazon RDS for PostgreSQL","PostgreSQLデータベースの運用を支えるAWSのマネージドサービスです。",{"id":87,"label":88,"shortDefinition":89,"role":65},"hallucination","幻覚","生成AIが、根拠のない内容をもっともらしく出力する現象です。",{"id":91,"label":92,"shortDefinition":93,"role":65},"evaluation","評価","定めたデータと基準で、モデルやアプリの品質・安全性・価値を測る活動です。",[95,100,112,121,130,136,142,148,164,173],{"id":96,"type":97,"title":7,"subtitle":98,"shortDefinition":99},"hero","HeroBlock","外部知識を検索し、根拠とともにモデルへ渡す","質問に関連する外部資料を検索し、その内容を文脈として基盤モデルへ渡して応答を生成する構成です。",{"id":101,"type":102,"term":103,"definition":104,"plainExplanation":105,"presentation":106,"supportLabel":107,"keywords":108},"definition","DefinitionBlock","Retrieval Augmented Generation（RAG）","[[cloze:retrieval-augmented-generation|検索拡張生成（RAG）]]とは、関連する外部資料を検索し、その内容を基盤モデルへ渡して応答を生成する方式です。","資料を[[term:chunking|チャンク化]]し、[[term:embedding|埋め込み]]を[[term:vector-database|ベクトルデータベース]]へ保存します。質問時は[[term:semantic-search|意味検索]]で関連部分を取得します。[[term:amazon-bedrock-knowledge-bases|Amazon Bedrock Knowledge Bases]]は取得と生成の統合を支援します。","natural","言葉を分けると",[109,110,111],"外部資料","意味的な近さ","根拠付き応答",{"id":113,"type":114,"steps":115,"title":120},"decision-steps","StepBlock",[116,117,118,119],"信頼できる資料を収集し、意味のまとまりでチャンク化する。","チャンクの埋め込みを[[term:vector|ベクトル]]として、出典情報と一緒に検索基盤へ保存する。","質問に近いチャンクを取得し、質問と一緒に[[term:foundation-model|FM]]へ渡す。","応答と引用が取得資料に支えられるか評価する。","RAGの処理順序",{"id":122,"type":123,"title":124,"points":125,"body":129},"exam-cues","KeyPointBlock","判断で押さえる境界",[126,127,128],"RAGは検索した外部知識を文脈へ追加","ベクトルストアは埋め込みと出典を管理","検索品質と生成品質を別々に評価","RAGは最新資料を参照しやすくしますが、[[term:hallucination|幻覚]]を完全には消しません。保存候補には[[term:amazon-opensearch-service|Amazon OpenSearch Service]]、[[term:amazon-aurora|Amazon Aurora]]、[[term:amazon-neptune|Amazon Neptune]]、[[term:amazon-rds-for-postgresql|Amazon RDS for PostgreSQL]]があります。検索、権限、出典、生成を分けて[[term:evaluation|評価]]します。",{"id":131,"type":132,"src":133,"alt":134,"caption":135},"diagram","DiagramBlock","\u002Fassets\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications\u002Frag-and-vector-databases\u002Frag-and-vector-databases.svg","RAGとベクトルデータベースについて資料→チャンク・埋め込み、質問→類似検索、関連資料を取得、FMが根拠付き応答の関係を文字ラベルと矢印で示した図","資料登録と質問時検索が合流し、FMへ根拠を渡す流れです。検索失敗と生成失敗を別に追跡します。",{"id":137,"type":138,"scenario":139,"explanation":140,"result":141},"worked-example","ExampleBlock","毎週改定される社内旅費規程へ質問できるアプリを作ります。","規程を版と部署権限付きでチャンク化し、埋め込みと原文URLを保存します。質問時に適用部署と日付で絞って類似検索し、取得文と出典をFMへ渡します。回答は引用箇所と版を表示し、該当根拠がなければ断定しない設計にします。","モデル再学習を待たず規程更新を反映し、回答根拠を追える構成になります。",{"id":36,"type":143,"warningTitle":144,"message":145,"severity":146,"action":147},"WarningBlock","RAGは幻覚を完全には消さない","無関係な資料を取得したり、モデルが資料を誤読したり、根拠外を補って生成したりできます。取得結果と最終回答の双方を評価します。","warning","回答ごとに、取得資料、出典、適用版、回答中の主張が対応しているか確認します。",{"id":149,"type":150,"title":151,"questions":152},"quiz","QuizBlock","AIF-C01 判断チェック",[153],{"question":154,"choices":155,"answerIndex":159,"choiceExplanations":160},"RAGでベクトル検索が担う役割はどれですか。",[156,157,158],"質問に意味的に近い資料片を取得する","基盤モデルの重みを毎回再学習する","出力の正しさを無条件に保証する",0,[161,162,163],"正解です。関連文脈を選びます。","RAGは通常、回答時に外部知識を追加します。","検索と生成の誤りは残ります。",{"id":165,"type":166,"title":167,"points":168},"summary","SummaryBlock","まとめ",[169,170,171,172],"RAGは検索した外部知識で生成を補強","資料をチャンク化し埋め込みを保存","AWSには複数のベクトル保存候補がある","検索・権限・出典・生成を別々に評価する",{"id":174,"type":175,"title":176,"nextTopics":177,"relatedKeywords":179},"next-action","NextActionBlock","次は「基盤モデルのカスタマイズ戦略」へ",[178],"it-aws-aws-certified-ai-practitioner-foundation-model-applications-foundation-model-customization-strategies",[180,20,181],"プロンプト","微調整",{"path":10,"title":19,"description":183,"parentPath":184,"accentToken":185},"基盤モデルの選択、RAG、プロンプト、カスタマイズ、エージェント、評価をAIF-C01の設計判断として整理します。","\u002Fit\u002Faws\u002Faws-certified-ai-practitioner","cloud-amber",{"title":187,"description":188,"canonicalUrl":189,"ogImage":190},"RAGとベクトルデータベース | AWS Certified AI Practitioner","AIF-C01対策としてRAGとベクトルデータベースの定義、判断軸、具体例、誤解しやすい境界を公式試験範囲に沿って学びます。","https:\u002F\u002Fzeqnilo.com\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications\u002Frag-and-vector-databases","https:\u002F\u002Fzeqnilo.com\u002Fassets\u002Fsite\u002Fdefault-og.svg",[192,195,198,199,200],{"title":193,"path":194},"IT","\u002Fit",{"title":196,"path":197},"AWS","\u002Fit\u002Faws",{"title":17,"path":184},{"title":19,"path":10},{"title":7,"path":9},[],[203],{"id":178,"title":204,"summary":205,"canonicalPath":206,"categoryPath":10,"categoryTitle":19,"difficulty":11,"targetAudience":12,"estimatedMinutes":13,"publishedAt":14,"updatedAt":14,"reviewStatus":15,"tags":207},"基盤モデルのカスタマイズ戦略","基盤モデルのカスタマイズ戦略を、AIF-C01の公式objective 3.1.5に沿って、用語、判断順序、例、誤解修正で整理します。","\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications\u002Ffoundation-model-customization-strategies",[17,18,19,180,20,181,208,209],"継続事前学習","モデル蒸留",[211],{"id":212,"title":213,"summary":214,"canonicalPath":215,"categoryPath":10,"categoryTitle":19,"difficulty":11,"targetAudience":12,"estimatedMinutes":13,"publishedAt":14,"updatedAt":14,"reviewStatus":15,"tags":216},"it-aws-aws-certified-ai-practitioner-foundation-model-applications-foundation-model-selection-and-inference-parameters","基盤モデル選択と推論パラメータ","基盤モデル選択と推論パラメータを、AIF-C01の公式objective 3.1.1・3.1.2に沿って、用語、判断順序、例、誤解修正で整理します。","\u002Fit\u002Faws\u002Faws-certified-ai-practitioner\u002Ffoundation-model-applications\u002Ffoundation-model-selection-and-inference-parameters",[17,18,19,217,218,219,220,221],"基盤モデル選択","temperature","入力長","出力長","プロンプトキャッシュ",1787370539935]