RAGについて情報をまとめる
RAGについてまとめ
RAG情報が溢れているので整理しています。
RAGの概要・入門
https://speakerdeck.com/smiyawaki0820/retrieval-based-lm-rag-system-zatukurili-jie-suru
https://zenn.dev/sompojapan_dx/articles/eb755a18e893ce
https://note.com/qunasys/n/n189f930059a6
https://www.deeplearning.ai/short-courses/knowledge-graphs-rag/
https://internet.watch.impress.co.jp/docs/column/shimizu/1580947.html
https://tech-lab.sios.jp/archives/40774
https://www.docswell.com/s/hoxo-m_inc/K4V3MW-rag
RAGの性能改善のテクニック
まとめ
https://pub.towardsai.net/advanced-rag-techniques-an-illustrated-overview-04d193d8fec6
https://www.jiang.jp/posts/20230611_retreiver/
https://speakerdeck.com/mkazutaka/gptsyorijing-du-nogao-iragsisutemunogou-zhu
https://zenn.dev/knowledgesense/articles/47de9ead8029ba
https://dev.classmethod.jp/articles/rag-knowledge-on-real-projects/
https://tech.dentsusoken.com/entry/parameter_tuning_for_rag
https://github.com/RManLuo/Awesome-LLM-KG
https://qiita.com/jw-automation/items/045917be7b558509fdf2
https://sue124.hatenablog.com/entry/2024/07/02/233616
https://zenn.dev/minedia/articles/8f4ef7f2daed11
https://zenn.dev/umi_mori/books/llm-rag-langchain-python
手法
https://note.com/npaka/n/n0be34a51395d
https://speakerdeck.com/mkazutaka/gptsyorijing-du-nogao-iragsisutemunogou-zhu
https://nikkie-ftnext.hatenablog.com/entry/grasp-hyde-hypothetical-document-embeddings
https://hironsan.hatenablog.com/entry/information-retrieval-with-reranker
https://docs.llamaindex.ai/en/stable/examples/index_structs/doc_summary/DocSummary/
https://note.com/npaka/n/n0be34a51395d
https://tech-lab.sios.jp/archives/38900
https://note.com/qunasys/n/nf9ee9a4e5d60
https://hironsan.hatenablog.com/entry/rag-with-knowledge-graph
https://zenn.dev/knowledgesense/articles/bb5e15abb3c547
https://zenn.dev/knowledgesense/articles/67dd2a41fc4d0b
https://zenn.dev/knowledgesense/articles/508187f1c616e3
https://zenn.dev/knowledgesense/articles/8c23c35fa715c9
https://qiita.com/kernelian/items/18f308332880a73b929a
https://zenn.dev/knowledgesense/articles/913d07f490e9c7
https://zenn.dev/knowledgesense/articles/4017fe6398b54e
https://www.tech-street.jp/entry/2024/06/26/134337
https://zenn.dev/knowledgesense/articles/110a81646806e6
https://zenn.dev/knowledgesense/articles/c9c50ae37f2ebf
https://zenn.dev/knowledgesense/articles/7fc387fc132c76
https://zenn.dev/knowledgesense/articles/653229e5c37f4e
https://note.com/npaka/n/n14dbea326178
https://acro-engineer.hatenablog.com/entry/2024/09/18/120000
https://zenn.dev/knowledgesense/articles/f84fab70ce04de
https://zenn.dev/knowledgesense/articles/1ecd331dc6b589
https://zenn.dev/knowledgesense/articles/e0ade68c265200
https://zenn.dev/knowledgesense/articles/e35011933152e2
RAG関係の論文
RAG関係のサーベイ論文
https://arxiv.org/abs/2312.10997
画像はRetrieval-Augmented Generation for Large Language Models: A Surveyより引用
時系列のまとめ
https://note.com/rrrrrrrrrr_666/n/nb8c9010ce603
まとめのGitHubリポジトリ
https://github.com/Tongji-KGLLM/RAG-Survey
サーベイ論文の解説記事 RAG（検索拡張生成）包括的な論文をわかりやすく解説
コサイン類似度が本当に適しているのかをといかける論文
https://arxiv.org/abs/2403.05440
retrieval-augmented thoughts(RAT)という手法について書かれた論文
https://arxiv.org/abs/2403.05313
RAGのエラーの分類に関する論文
https://arxiv.org/abs/2403.01193
HyDEという手法の論文
https://arxiv.org/abs/2212.10496
メタ認知をRAGに適用
https://arxiv.org/abs/2402.11626
Self RAG
https://arxiv.org/abs/2310.11511
NVIDIA Order-Preserve RAG
https://arxiv.org/abs/2409.01666
RAGのスケーリング則
https://arxiv.org/abs/2410.04343
グラフ（GraphRAG）
https://www.creationline.com/tech-blog/68014
https://qiita.com/Ruuchami/items/294d797798d4fca346ad
https://arxiv.org/abs/2306.08302
https://arxiv.org/abs/2402.05391
https://arxiv.org/abs/2408.04913
https://pnch.hatenablog.com/entry/graphrag
https://www.alpha.co.jp/blog/202408_01/
https://zenn.dev/knowledgesense/articles/077ad1ab0f9ff6
https://techblog.insightedge.jp/entry/knowledgegraph-recommendation
https://qiita.com/ssfujita/items/65a952f299190f4c1e6a
https://zenn.dev/kun432/scraps/1f28e5d20dfdf5
https://github.com/gusye1234/nano-graphrag
https://github.com/HKUDS/LightRAG
実践
https://note.com/npaka/n/n6d33c2181050
https://note.com/npaka/n/n63afe0e6684a
https://zenn.dev/cloud_ace/articles/19bd3554ac8432
https://zenn.dev/spiralai/articles/8af7cbf526c2e1
https://tech-blog.abeja.asia/entry/in-house-jargon-slackbot-with-rag-202402
https://zenn.dev/smartshopping/articles/rag-with-bedrock-and-openai
https://qiita.com/ksonoda/items/ba6d7b913fc744db3d79
https://qiita.com/mitsumizo/items/469d79c5e81d9189a9e4
https://speakerdeck.com/soracom/history-of-soracom-support-bot-20240321
https://qiita.com/minorun365/items/24dfb0ea3afde6ed0a56
https://qiita.com/moritalous/items/76ba9f2ad200df335d07
https://zenn.dev/yumefuku/articles/llm-langchain-rag
https://qiita.com/Naoki_Ishihara/items/662d70a9bd0dc3a8c9ce?1
https://zenn.dev/aidemy/articles/97d5fb6ac03a4f
https://zenn.dev/umi_mori/books/prompt-engineer/viewer/pdf_langchain_chatgpt
https://qiita.com/hiroki_okuhata_int/items/7102bab7d96eb2574e7d
https://zenn.dev/aidemy/articles/cd79fe964ebbff
https://blog.g-gen.co.jp/entry/comparing-rag-architecture-across-cloud-vendors
https://acro-engineer.hatenablog.com/entry/2024/08/22/120000
https://speakerdeck.com/forrep/rag-does-not-need-a-vector-db
https://blog.studysapuri.jp/entry/2024/06/26/100000
https://note.com/kan_hatakeyama/n/nd851fd842515
https://techblog.exawizards.com/entry/2024/11/01/151609
評価
RAGの評価ソフト
https://github.com/explodinggradients/ragas
https://www.nogawanogawa.com/entry/ragas
RAGの評価に関するサーベイ論文
https://arxiv.org/abs/2405.07437
https://zenn.dev/knowledgesense/articles/bfd100c51ff34e
https://www.nogawanogawa.com/entry/rag_eval
https://github.com/IntelLabs/RAGFoundry
https://techblog.lycorp.co.jp/ja/20240819a
https://tech.assured.jp/entry/2024/08/20/130000
https://zenn.dev/knowledgesense/articles/90ac35eedf8b7c
https://www.nogawanogawa.com/entry/rag_eval_2
ブログ記事
https://note.com/npaka/n/n53e8aabed0f2
https://note.com/npaka/n/n62cd25213679
https://logmi.jp/tech/articles/329143
LLMのRAGを用いたコンペ
https://zenn.dev/nishimoto/articles/aff1fba9c75c34
https://docs.llamaindex.ai/en/stable/examples/index_structs/doc_summary/DocSummary/
https://blog.soracom.com/ja-jp/2024/07/12/soracom-support-bot/
https://zenn.dev/google_cloud_jp/articles/598d52341cc56f
https://zenn.dev/kun432/scraps/c0642e79169635
https://qiita.com/ksonoda/items/f3c703fb7b689fd65868
https://zenn.dev/minedia/articles/d9f01aa05bc880
まとめ
随時更新中です。
関連記事
https://zenn.dev/karaage0703/articles/c8baa66c40f9b7
https://zenn.dev/karaage0703/articles/90d4de4596b262
Discussion