{"id":4248,"date":"2024-01-08T05:17:00","date_gmt":"2024-01-07T20:17:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4248"},"modified":"2024-01-08T05:17:00","modified_gmt":"2024-01-07T20:17:00","slug":"docllm","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4248","title":{"rendered":"DocLLM"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>DocLLM: A layout-aware generative language model for multimodal document understanding\u00a0<\/strong>[12.1]<br>\u672c\u7a3f\u3067\u306f,\u5f93\u6765\u306e\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb(LLM)\u306e\u8efd\u91cf\u62e1\u5f35\u3067\u3042\u308bDocLLM\u306b\u3064\u3044\u3066\u8ff0\u3079\u308b\u3002 \u672c\u30e2\u30c7\u30eb\u306f,\u7a7a\u9593\u914d\u7f6e\u69cb\u9020\u3092\u7d44\u307f\u8fbc\u3080\u305f\u3081\u306e\u5883\u754c\u30dc\u30c3\u30af\u30b9\u60c5\u5831\u306b\u306e\u307f\u7126\u70b9\u3092\u3042\u3066\u308b\u3002 \u6211\u3005\u306e\u30bd\u30ea\u30e5\u30fc\u30b7\u30e7\u30f3\u306f\u3001\u3059\u3079\u3066\u306e\u30bf\u30b9\u30af\u306b\u307e\u305f\u304c\u308b16\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306e\u3046\u306114\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067SotA LLM\u3088\u308a\u3082\u512a\u308c\u3066\u304a\u308a\u3001\u3053\u308c\u307e\u3067\u898b\u3064\u304b\u3089\u306a\u304b\u3063\u305f5\u3064\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306e\u3046\u30614\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067\u5341\u5206\u306b\u4e00\u822c\u5316\u3055\u308c\u3066\u3044\u308b\u3053\u3068\u3092\u5b9f\u8a3c\u3057\u3066\u3044\u307e\u3059\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2401.00908v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2401.00908v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Sun, 31 Dec 2023 22:37:52 GMT)<\/li>\n\n\n\n<li>bounding box\u306e\u60c5\u5831\u3092\u7d44\u307f\u8fbc\u3093\u3060LLM\u3001\u753b\u50cf\u3078\u306e\u62e1\u5f35\u3088\u308a\u3082\u52b9\u7387\u7684\u3068\u306e\u3053\u3068\u3002\u5b9f\u88c5\u4e0a\u6709\u7528\u306a\u30a2\u30d7\u30ed\u30fc\u30c1\u306b\u601d\u3048\u308b\u3002\u8457\u8005\u304cJPMorgan AI Research\u3068\u3044\u3046\u306e\u3082\u8208\u5473\u6df1\u3044\u3002<\/li>\n\n\n\n<li>\u300cDocLLM is a multi-modal system that integrates lightweight visual information by utilizing the spatial positions and dimensions of text tokens obtained using OCR.\u300d\u3068\u3044\u3046\u3053\u3068\u3067bounding box\u306fOCR\u304b\u3089\u5f97\u308b\u306e\u304c\u524d\u63d0\u3067\u306f\u3042\u308b\u304c\u3001\u30c6\u30ad\u30b9\u30c8\u3084\u30d6\u30ed\u30c3\u30af\u69cb\u9020\u304c\u5f97\u3089\u308c\u308b\u96fb\u5b50\u30d5\u30a1\u30a4\u30eb\u304c\u4f7f\u3048\u308b\u5834\u5408\u306f\u3055\u3089\u306b\u3088\u304f\u52d5\u304d\u305d\u3046\uff08\u975e\u73fe\u5b9f\u7684\u306a\u4eee\u5b9a\u3067\u3082\u306a\u3044\uff09\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[117,223,524],"class_list":["post-4248","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-document-understanding","tag-llm","tag-524"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4248","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=4248"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4248\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4248"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4248"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4248"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}