$\"image.png\"$

\n","updatedAt":"2025-02-04T09:35:57.160Z","author":{"_id":"643407dd4b34368fdb0149e8","avatarUrl":"/avatars/9477b9267d5692a4fe59e30590e9639d.svg","fullname":"Xinyan Guan","name":"xinyan233333","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.40066224336624146},"editors":["xinyan233333"],"editorAvatarUrls":["/avatars/9477b9267d5692a4fe59e30590e9639d.svg"],"reactions":[],"isReport":false}},{"id":"67a2bfdc0995c2bd768c9e1a","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false},"createdAt":"2025-02-05T01:33:16.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [SUGAR: Leveraging Contextual Confidence for Smarter Retrieval](https://huggingface.co/papers/2501.04899) (2025)\n* [RAG-Star: Enhancing Deliberative Reasoning with Retrieval Augmented Verification and Refinement](https://huggingface.co/papers/2412.12881) (2024)\n* [RPO: Retrieval Preference Optimization for Robust Retrieval-Augmented Generation](https://huggingface.co/papers/2501.13726) (2025)\n* [RbFT: Robust Fine-tuning for Retrieval-Augmented Generation against Retrieval Defects](https://huggingface.co/papers/2501.18365) (2025)\n* [To Retrieve or Not to Retrieve? Uncertainty Detection for Dynamic Retrieval Augmented Generation](https://huggingface.co/papers/2501.09292) (2025)\n* [Retrieval-Augmented Semantic Parsing: Using Large Language Models to Improve Generalization](https://huggingface.co/papers/2412.10207) (2024)\n* [Parametric Retrieval Augmented Generation](https://huggingface.co/papers/2501.15915) (2025)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2025-02-05T01:33:16.342Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6936085820198059},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2502.01142","authors":[{"_id":"67a1b4630e9634919de9bc52","user":{"_id":"643407dd4b34368fdb0149e8","avatarUrl":"/avatars/9477b9267d5692a4fe59e30590e9639d.svg","isPro":false,"fullname":"Xinyan Guan","user":"xinyan233333","type":"user"},"name":"Xinyan Guan","status":"claimed_verified","statusLastChangedAt":"2025-02-04T09:39:08.849Z","hidden":false},{"_id":"67a1b4630e9634919de9bc53","user":{"_id":"657bef0fff16eeb2ee40ed9c","avatarUrl":"/avatars/2a436b1d9b04c611a795f10363150aca.svg","isPro":false,"fullname":"zeng","user":"zengjiali","type":"user"},"name":"Jiali Zeng","status":"admin_assigned","statusLastChangedAt":"2025-02-04T17:36:59.186Z","hidden":false},{"_id":"67a1b4630e9634919de9bc54","user":{"_id":"64cb254871a7bbb60c17d5fa","avatarUrl":"/avatars/5121fd5b7b55d275eba3947f3f4c034d.svg","isPro":false,"fullname":"Fandong Meng","user":"fandong","type":"user"},"name":"Fandong Meng","status":"admin_assigned","statusLastChangedAt":"2025-02-04T17:36:44.284Z","hidden":false},{"_id":"67a1b4630e9634919de9bc55","user":{"_id":"667b74b9279199a7c610687f","avatarUrl":"/avatars/9834e9971579655a4c387a306c610f57.svg","isPro":false,"fullname":"Chunlei Xin","user":"meow77","type":"user"},"name":"Chunlei Xin","status":"admin_assigned","statusLastChangedAt":"2025-02-04T17:37:04.877Z","hidden":false},{"_id":"67a1b4630e9634919de9bc56","user":{"_id":"6216496a9b34d2fb49144599","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6216496a9b34d2fb49144599/41CKA_h1Ffj3RzVabSAkm.jpeg","isPro":false,"fullname":"Yaojie Lu","user":"luyaojie","type":"user"},"name":"Yaojie Lu","status":"admin_assigned","statusLastChangedAt":"2025-02-04T17:37:11.297Z","hidden":false},{"_id":"67a1b4630e9634919de9bc57","user":{"_id":"6711c702f858a456b4b9f3a4","avatarUrl":"/avatars/178e9567c3111ab22717c3c0dd003a6a.svg","isPro":false,"fullname":"Hongyu Lin","user":"sanmusunrise","type":"user"},"name":"Hongyu Lin","status":"admin_assigned","statusLastChangedAt":"2025-02-04T17:37:20.338Z","hidden":false},{"_id":"67a1b4630e9634919de9bc58","user":{"_id":"65e99a77e71555ed193609cf","avatarUrl":"/avatars/38ceb127883944677665da967d17dd18.svg","isPro":false,"fullname":"Xianpei Han","user":"xphan","type":"user"},"name":"Xianpei Han","status":"admin_assigned","statusLastChangedAt":"2025-02-04T17:37:26.589Z","hidden":false},{"_id":"67a1b4630e9634919de9bc59","name":"Le Sun","hidden":false},{"_id":"67a1b4630e9634919de9bc5a","name":"Jie Zhou","hidden":false}],"publishedAt":"2025-02-03T08:22:45.000Z","submittedOnDailyAt":"2025-02-04T07:05:57.149Z","title":"DeepRAG: Thinking to Retrieval Step by Step for Large Language Models","submittedOnDailyBy":{"_id":"643407dd4b34368fdb0149e8","avatarUrl":"/avatars/9477b9267d5692a4fe59e30590e9639d.svg","isPro":false,"fullname":"Xinyan Guan","user":"xinyan233333","type":"user"},"summary":"Large Language Models (LLMs) have shown remarkable potential in reasoning\nwhile they still suffer from severe factual hallucinations due to timeliness,\naccuracy, and coverage of parametric knowledge. Meanwhile, integrating\nreasoning with retrieval-augmented generation (RAG) remains challenging due to\nineffective task decomposition and redundant retrieval, which can introduce\nnoise and degrade response quality. In this paper, we propose DeepRAG, a\nframework that models retrieval-augmented reasoning as a Markov Decision\nProcess (MDP), enabling strategic and adaptive retrieval. By iteratively\ndecomposing queries, DeepRAG dynamically determines whether to retrieve\nexternal knowledge or rely on parametric reasoning at each step. Experiments\nshow that DeepRAG improves retrieval efficiency while improving answer accuracy\nby 21.99%, demonstrating its effectiveness in optimizing retrieval-augmented\nreasoning.","upvotes":24,"discussionId":"67a1b4640e9634919de9bc8b","ai_summary":"DeepRAG improves answer accuracy by 21.99% through strategic and adaptive retrieval, modeled as a Markov Decision Process for efficient reasoning-augmented generation.","ai_keywords":["Markov Decision Process","retrieval-augmented reasoning","retrieval-augmented generation","parametric reasoning","task decomposition","strategic retrieval","adaptive retrieval"]},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"65d9903fdceb54d42011a98d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65d9903fdceb54d42011a98d/5jnLeCY9sDtS98JyO9qzX.jpeg","isPro":false,"fullname":"meng shao","user":"meng-shao","type":"user"},{"_id":"643407dd4b34368fdb0149e8","avatarUrl":"/avatars/9477b9267d5692a4fe59e30590e9639d.svg","isPro":false,"fullname":"Xinyan Guan","user":"xinyan233333","type":"user"},{"_id":"643b62ac065961b2252abb7a","avatarUrl":"/avatars/c7fb4d11f0d795a52bdc771c04a69a20.svg","isPro":false,"fullname":"zuijiang","user":"zuijiang","type":"user"},{"_id":"6393200e7a88f87c10056c43","avatarUrl":"/avatars/4ddde5cc4e5780e55d12ce0129234365.svg","isPro":false,"fullname":"zhongtianyun","user":"tianyumyum","type":"user"},{"_id":"66d0d88d3c5bc37ee07b937c","avatarUrl":"/avatars/6d50b37d00510e560d6c79ea346a0a5e.svg","isPro":false,"fullname":"Christophe Amoussouvi","user":"NeuralDev","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"648eb1eb59c4e5c87dc116e0","avatarUrl":"/avatars/c636cea39c2c0937f01398c94ead5dad.svg","isPro":false,"fullname":"fdsqefsgergd","user":"T-representer","type":"user"},{"_id":"6270324ebecab9e2dcf245de","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6270324ebecab9e2dcf245de/cMbtWSasyNlYc9hvsEEzt.jpeg","isPro":false,"fullname":"Kye Gomez","user":"kye","type":"user"},{"_id":"631e14ac473a6825f285e89d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/631e14ac473a6825f285e89d/K-6QnoeGLg8XFvbTMMdqA.jpeg","isPro":false,"fullname":"Yury Panikov","user":"panikov","type":"user"},{"_id":"64c1c77c245c55a21c6f5a13","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64c1c77c245c55a21c6f5a13/d9zlSksf3TxWpBbb-r0fd.jpeg","isPro":false,"fullname":"Reza Sayar","user":"Reza2kn","type":"user"},{"_id":"65e99a77e71555ed193609cf","avatarUrl":"/avatars/38ceb127883944677665da967d17dd18.svg","isPro":false,"fullname":"Xianpei Han","user":"xphan","type":"user"},{"_id":"60078446e55258e41786a959","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/60078446e55258e41786a959/UGPCE4YqG9BVMSf0YauxL.png","isPro":false,"fullname":"Motoki Wu","user":"tokestermw","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0}">

Papers

arxiv:2502.01142

DeepRAG: Thinking to Retrieval Step by Step for Large Language Models

Published on Feb 3, 2025

· Submitted by

Xinyan Guan on Feb 4, 2025

Upvote

Authors:

Xinyan Guan ,

Jiali Zeng ,

Fandong Meng ,

Chunlei Xin ,

Yaojie Lu ,

Hongyu Lin ,

Xianpei Han ,

Abstract

DeepRAG improves answer accuracy by 21.99% through strategic and adaptive retrieval, modeled as a Markov Decision Process for efficient reasoning-augmented generation.

AI-generated summary

Large Language Models (LLMs) have shown remarkable potential in reasoning while they still suffer from severe factual hallucinations due to timeliness, accuracy, and coverage of parametric knowledge. Meanwhile, integrating reasoning with retrieval-augmented generation (RAG) remains challenging due to ineffective task decomposition and redundant retrieval, which can introduce noise and degrade response quality. In this paper, we propose DeepRAG, a framework that models retrieval-augmented reasoning as a Markov Decision Process (MDP), enabling strategic and adaptive retrieval. By iteratively decomposing queries, DeepRAG dynamically determines whether to retrieve external knowledge or rely on parametric reasoning at each step. Experiments show that DeepRAG improves retrieval efficiency while improving answer accuracy by 21.99%, demonstrating its effectiveness in optimizing retrieval-augmented reasoning.

View arXiv page View PDF Add to collection