Historical documents act as invaluable knowledge archives but often suffer from illegibility due to physical deterioration and damage. While existing restoration methods based on masked language modeling effectively utilize local context, they struggle to restore named entities that require external historical knowledge. To address this limitation, we introduce a novel framework for historical document restoration that leverages large language models with retrieval-augmented generation (RAG). By combining the implicit knowledge of pretrained LLMs with explicitly retrieved external context, our model ARI effectively mitigates the challenge of inferring context-dependent proper nouns. Extensive experiments on Korean historical documents demonstrate that our approach significantly outperforms baselines, achieving substantial gains in restoring both general characters and named entities. Furthermore, comprehensive evaluations including expert assessments confirm that ARI serves as a practical tool for domain experts, promising to accelerate the analysis of historical records.<br><a href=\"https://cdn-uploads.huggingface.co/production/uploads/64c4f93cfafa16b5144ad824/xFT4QiZRXPhcu_pdL4P3K.jpeg\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/64c4f93cfafa16b5144ad824/xFT4QiZRXPhcu_pdL4P3K.jpeg\" alt=\"main_performance\"></a></p>\n","updatedAt":"2026-07-28T03:41:07.282Z","author":{"_id":"64c4f93cfafa16b5144ad824","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64c4f93cfafa16b5144ad824/2i1SX8pqS9ESERx0xTTBB.png","fullname":"Gabeen Kim","name":"EvelynKimmy","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8862141370773315},"editors":["EvelynKimmy"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/64c4f93cfafa16b5144ad824/2i1SX8pqS9ESERx0xTTBB.png"],"reactions":[{"reaction":"🚀","users":["kyeongpil","HiMyNameIsDavidKim","j5nim"],"count":3},{"reaction":"🔥","users":["kyeongpil","HiMyNameIsDavidKim","j5nim"],"count":3},{"reaction":"❤️","users":["kyeongpil","HiMyNameIsDavidKim","j5nim"],"count":3},{"reaction":"🤗","users":["kyeongpil","HiMyNameIsDavidKim","j5nim"],"count":3}],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2607.21936","authors":[{"_id":"6a66a7c6ab9cdf9be57949fb","user":{"_id":"64c4f93cfafa16b5144ad824","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64c4f93cfafa16b5144ad824/2i1SX8pqS9ESERx0xTTBB.png","isPro":false,"fullname":"Gabeen Kim","user":"EvelynKimmy","type":"user","name":"EvelynKimmy"},"name":"Gabeen Kim","status":"claimed_verified","statusLastChangedAt":"2026-07-27T00:45:04.577Z","hidden":false},{"_id":"6a66a7c6ab9cdf9be57949fc","name":"Kyeongpil Kang","hidden":false}],"publishedAt":"2026-07-24T00:00:00.000Z","submittedOnDailyAt":"2026-07-28T00:00:00.000Z","title":"Leveraging External Knowledge for Historical Document Restoration via Retrieval-Augmented Large Language Models","submittedOnDailyBy":{"_id":"64c4f93cfafa16b5144ad824","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64c4f93cfafa16b5144ad824/2i1SX8pqS9ESERx0xTTBB.png","isPro":false,"fullname":"Gabeen Kim","user":"EvelynKimmy","type":"user","name":"EvelynKimmy"},"summary":"Historical documents act as invaluable knowledge archives but often suffer from illegibility due to physical deterioration and damage. While existing restoration methods based on masked language modeling effectively utilize local context, they struggle to restore named entities that require external historical knowledge. To address this limitation, we introduce a novel framework for historical document restoration that leverages large language models with retrieval-augmented generation (RAG). By combining the implicit knowledge of pre-trained LLMs with explicitly retrieved external context, our model ARI effectively mitigates the challenge of inferring context-dependent proper nouns. Extensive experiments on Korean historical documents demonstrate that our approach significantly outperforms baselines, achieving substantial gains in restoring both general characters and named entities. Furthermore, comprehensive evaluations including expert assessments confirm that ARI serves as a practical tool for domain experts, promising to accelerate the analysis of historical records.","upvotes":3,"discussionId":"6a66a7c6ab9cdf9be57949fd","githubRepo":"https://github.com/EvelynKimm/ARI","githubRepoAddedBy":"user","githubStars":1,"organization":{"_id":"6a0136095408a27490d2f010","name":"KNU-Korea","fullname":"Kangwon National University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/6a01344cffe530021049094b/gkM-Tx35hgLIKHydY5Zgp.jpeg"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"631c386bc73939ffc0716a37","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1662793811119-noauth.jpeg","isPro":false,"fullname":"SeongWan Kim","user":"idgmatrix","type":"user"},{"_id":"63bbd400b8c61b8aa492f961","avatarUrl":"/avatars/4c43cdd1352028296da54f57d52ee35e.svg","isPro":false,"fullname":"Kyeongpil Kang","user":"kyeongpil","type":"user"},{"_id":"69b902579d2809a871239174","avatarUrl":"/avatars/34b0d89a159c61e7b6a702870f8bb4ff.svg","isPro":false,"fullname":"JEONGHYEON LEE","user":"j5nim","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6a0136095408a27490d2f010","name":"KNU-Korea","fullname":"Kangwon National University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/6a01344cffe530021049094b/gkM-Tx35hgLIKHydY5Zgp.jpeg"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2607/2607.21936.md","query":{}}">
Leveraging External Knowledge for Historical Document Restoration via Retrieval-Augmented Large Language Models
Abstract
Historical documents act as invaluable knowledge archives but often suffer from illegibility due to physical deterioration and damage. While existing restoration methods based on masked language modeling effectively utilize local context, they struggle to restore named entities that require external historical knowledge. To address this limitation, we introduce a novel framework for historical document restoration that leverages large language models with retrieval-augmented generation (RAG). By combining the implicit knowledge of pre-trained LLMs with explicitly retrieved external context, our model ARI effectively mitigates the challenge of inferring context-dependent proper nouns. Extensive experiments on Korean historical documents demonstrate that our approach significantly outperforms baselines, achieving substantial gains in restoring both general characters and named entities. Furthermore, comprehensive evaluations including expert assessments confirm that ARI serves as a practical tool for domain experts, promising to accelerate the analysis of historical records.
Community
Historical documents act as invaluable knowledge archives but often suffer from illegibility due to physical deterioration and damage. While existing restoration methods based on masked language modeling effectively utilize local context, they struggle to restore named entities that require external historical knowledge. To address this limitation, we introduce a novel framework for historical document restoration that leverages large language models with retrieval-augmented generation (RAG). By combining the implicit knowledge of pretrained LLMs with explicitly retrieved external context, our model ARI effectively mitigates the challenge of inferring context-dependent proper nouns. Extensive experiments on Korean historical documents demonstrate that our approach significantly outperforms baselines, achieving substantial gains in restoring both general characters and named entities. Furthermore, comprehensive evaluations including expert assessments confirm that ARI serves as a practical tool for domain experts, promising to accelerate the analysis of historical records.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2607.21936 in a model README.md to link it from this page.
Cite arxiv.org/abs/2607.21936 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2607.21936 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.