{"id":3865,"date":"2023-09-25T07:20:00","date_gmt":"2023-09-24T22:20:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=3865"},"modified":"2023-09-25T07:20:00","modified_gmt":"2023-09-24T22:20:00","slug":"pdftriage","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=3865","title":{"rendered":"PDFTriage"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>PDFTriage: Question Answering over Long, Structured Documents&nbsp;<\/strong>[64.6]<br>\u69cb\u9020\u5316\u6587\u66f8\u3092\u30d7\u30ec\u30fc\u30f3\u30c6\u30ad\u30b9\u30c8\u3068\u3057\u3066\u8868\u73fe\u3059\u308b\u3053\u3068\u306f\u3001\u3053\u308c\u3089\u306e\u6587\u66f8\u3092\u30ea\u30c3\u30c1\u306a\u69cb\u9020\u3067\u30e6\u30fc\u30b6\u7cbe\u795e\u30e2\u30c7\u30eb\u3068\u77db\u76fe\u3059\u308b\u3002 \u672c\u7a3f\u3067\u306f,\u69cb\u9020\u3084\u5185\u5bb9\u306b\u57fa\u3065\u3044\u3066,\u30e2\u30c7\u30eb\u304c\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u3092\u691c\u7d22\u3067\u304d\u308bPDFTriage\u3092\u63d0\u6848\u3059\u308b\u3002 \u672c\u5b9f\u9a13\u306f,\u65e2\u5b58\u306e\u691c\u7d22\u62e1\u5f35LPM\u304c\u5931\u6557\u3059\u308b\u554f\u984c\u306b\u5bfe\u3057\u3066,\u63d0\u6848\u3057\u305fPDFTriage-augmented\u30e2\u30c7\u30eb\u306e\u6709\u52b9\u6027\u3092\u5b9f\u8a3c\u3059\u308b\u3082\u306e\u3067\u3042\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2309.08872v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2309.08872v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Sat, 16 Sep 2023 04:29:05 GMT)<\/li>\n\n\n\n<li>\u69cb\u9020\u5316\u6587\u66f8\u3078\u306b\u3064\u3044\u3066\u5206\u6790\u3057\u3001PDFTriage\uff08Generate document metadata \u2192 LLM-based triage \u2192 Answer using retrieved content\uff09\u3092\u63d0\u6848\u3001\u52b9\u679c\u3092\u78ba\u8a8d\u30022 step\u76ee\u304cfetch_pages, fetch_sections, fetch_table, fetch_figure, retrieve\u3092API\u7d4c\u7531\u3067\u547c\u3073\u51fa\u3059\u8a2d\u8a08\u306b\u306a\u3063\u3066\u3044\u308b\u306e\u304c\u9762\u767d\u3044\u3002<\/li>\n\n\n\n<li>\u300cWe identify a gap in question answering over structured documents with current LLM approaches, namely treating documents as plain text rather than structured objects;\u300d\u306f\u307e\u3055\u306b\u305d\u306e\u901a\u308a\u3060\u3068\u601d\u3046\u3002PDF Triage\u306e\u3088\u3046\u306a\u624b\u6cd5\u3067\u306f\u306a\u304fRAG\u306b\u8fd1\u5e74\u306eDocument understanding\u95a2\u9023\u306e\u6210\u679c\u3092\u53d6\u308a\u8fbc\u3080\u3088\u3046\u306a\u65b9\u5411\u6027\u3082\u3042\u308b\u3068\u601d\u3046\u304c\u3001\u65e9\u3081\u306b\u9032\u5316\u3057\u3066\u307b\u3057\u3044\u3068\u3053\u308d\u3002<\/li>\n<\/ul>\n\n\n\n<p>Document Understanding\u95a2\u9023\u3067\u3082LLM\u306e\u6d3b\u7528\u304c\u9032\u3080<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>LMDX: Language Model-based Document Information Extraction and Localization\u00a0<\/strong>[15.7]<br>\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb(LLM)\u306f\u81ea\u7136\u8a00\u8a9e\u51e6\u7406(NLP)\u306b\u9769\u547d\u3092\u3082\u305f\u3089\u3057\u305f \u672c\u7a3f\u3067\u306f,\u6587\u66f8\u60c5\u5831\u306e\u62bd\u51fa\u306b\u4efb\u610f\u306e LLM \u3092\u9069\u7528\u3059\u308b\u624b\u6cd5\u3067\u3042\u308b LMDX \u3092\u7d39\u4ecb\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2309.10952v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2309.10952v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 19 Sep 2023 22:32:56 GMT)<\/li>\n\n\n\n<li>LLM\u3092\u7528\u3044\u305f\u6587\u66f8\u304b\u3089\u306e\u60c5\u5831\u62bd\u51fa\uff08OCR\u304c\u8d77\u70b9\uff09\u3001LayoutLMv3\u3092\u8d85\u3048\u3066\u3044\u308b\u306e\u304c\u51c4\u3044\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Document Understanding\u95a2\u9023\u3067\u3082LLM\u306e\u6d3b\u7528\u304c\u9032\u3080<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[117,223],"class_list":["post-3865","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-document-understanding","tag-llm"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3865","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=3865"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3865\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3865"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3865"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3865"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}