{"id":5174,"date":"2026-05-18T11:54:18","date_gmt":"2026-05-18T03:54:18","guid":{"rendered":"https:\/\/www.15zhi.net\/blog\/?p=5174"},"modified":"2026-05-18T11:54:18","modified_gmt":"2026-05-18T03:54:18","slug":"202605%e8%ae%ba%e6%96%87%e7%a0%94%e8%af%bb-information-extraction-from-visually-rich-documents-using-llm-based-organization-of-documents-into-independent-textual-segments","status":"publish","type":"post","link":"https:\/\/www.15zhi.net\/blog\/202605%e8%ae%ba%e6%96%87%e7%a0%94%e8%af%bb-information-extraction-from-visually-rich-documents-using-llm-based-organization-of-documents-into-independent-textual-segments\/","title":{"rendered":"202605\u8bba\u6587\u7814\u8bfb-Information Extraction from Visually Rich Documents using LLM-based Organization of Documents into Independent Textual Segments"},"content":{"rendered":"\n

\u4f5c\u8005\uff1aAniket Bhattacharyya1, Anurag Tripathi, Ujjal Das, Archan Karmakar, Amit Pathak, Maneesh Gupta
\u6765\u6e90\uff1aACL 2025
\u65f6\u95f4\uff1a2025.5.18<\/p>\n\n\n\n

\u7814\u7a76\u80cc\u666f\u4e0e\u95ee\u9898<\/p>\n\n\n\n

Visually Rich Document Understanding (VRDU) VRDU \u5904\u7406\u7684\u662f\u540c\u65f6\u5305\u542b\u6587\u672c\u548c\u7248\u9762\u4fe1\u606f\u7684\u6587\u6863\u2014\u2014\u53d1\u7968\u3001\u8868\u5355\u3001\u5408\u540c\u3001\u6536\u636e\u7b49\u3002\u8fd9\u7c7b\u6587\u6863\u5728\u4f01\u4e1a\u4e2d\u91cf\u5927\u3001\u4ef7\u503c\u9ad8(\u5c24\u5176\u91d1\u878d\u3001\u6cd5\u5f8b\u573a\u666f),\u81ea\u52a8\u5316\u63d0\u53d6\u5173\u952e\u4fe1\u606f(KIE\/IE)\u662f\u957f\u671f\u7814\u7a76\u70ed\u70b9\u3002
\u5df2\u6709\u65b9\u6cd5\u6216\u8005\u4f1a\u63a8\u7406\u4f46\u770b\u4e0d\u61c2\u7248\u9762,\u6216\u8005\u61c2\u7248\u9762\u4f46\u4e0d\u4f1a\u63a8\u7406\u3002
\u4e09\u7c7b\u65e2\u6709\u65b9\u6cd5\u5404\u81ea\u7684\u74f6\u9888\uff1a<\/p>\n\n\n\n

\u65b9\u6cd5\u8def\u7ebf<\/td>\u4ee3\u8868\u5de5\u4f5c<\/td>\u4e3b\u8981\u7f3a\u9677<\/td><\/tr>
\u4f20\u7edf\u65b9\u6cd5<\/td>\u89c4\u5219\u3001RNN\u3001CNN\u3001Chargrid \u7b49<\/td>\u6a21\u677f\u9501\u6b7b\uff0c\u8fc1\u79fb\u6027\u5dee\uff0c\u9700\u8981\u5927\u91cf\u7ec4\u4ef6\u7ea7\u6807\u6ce8<\/td><\/tr>
Layout-aware NLP<\/td>LayoutLMv3, GeoLayoutLM, ERNIE-Layout, FormNetV2<\/td>\u672c\u8d28\u662f token \u5206\u7c7b\uff0c\u8981\u6c42\u7b54\u6848\u663e\u5f0f\u5b58\u5728\uff1b\u57fa\u51c6\u4e0a\u5f88\u5f3a\u4f46\u78b0\u5230\u65b0\u683c\u5f0f\u5c31\u5d29<\/td><\/tr>
LLM-based<\/td>DocLLM, LayoutLLM, LMDX<\/td>\u6709\u63a8\u7406\u4f46\u7f3a\u7248\u9762\u7406\u89e3\uff1bLMDX \u9700\u8981\u6807\u6ce8\u96c6\u4e2d\u6709\u540c\u683c\u5f0f\u6837\u672c\uff1b\u5728\u5f02\u6784\u57fa\u51c6\u4e0a\u8dd1\u4e0d\u8fc7 layout-aware \u65b9\u6cd5<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

\u8bba\u6587\u7ed9\u51fa\u7684”\u7406\u60f3 IE \u89e3\u51b3\u65b9\u6848”\u56db\u6761\u6807\u51c6(\u4e5f\u5373\u6587\u7ae0\u8d2f\u7a7f\u59cb\u7ec8\u7684\u8bc4\u4ef7 desiderata):
\u9ad8\u8d28\u91cf\u62bd\u53d6\u2014\u2014\u76ee\u6807\u5b9e\u4f53(\u516c\u53f8\u540d\u3001\u5730\u5740\u7b49)\u7684\u9ad8 P\/R
\u5904\u7406\u683c\u5f0f\u4e0e\u8bed\u8a00\u5f02\u6784\u6027\u2014\u2014\u540c\u4e00\u7cfb\u7edf\u8981\u80fd\u5904\u7406\u7f8e\u56fd\u6cd5\u5f8b\u4f20\u771f\u3001\u5370\u5c3c\u4fbf\u5229\u5e97\u53d1\u7968\u7b49\u5dee\u522b\u5de8\u5927\u7684\u6a21\u677f
\u5904\u7406\u65b0\u683c\u5f0f\u2014\u2014\u8bad\u7ec3\u9636\u6bb5\u672a\u89c1\u8fc7\u7684\u7248\u5f0f\u4e0d\u80fd\u76f4\u63a5\u5931\u8d25
\u652f\u6301 value-absent inference(\u9690\u542b\u503c\u63a8\u7406)\u2014\u2014\u6bd4\u5982”line item \u6570\u91cf”\u8fd9\u79cd\u6587\u6863\u91cc\u6ca1\u6709\u663e\u5f0f\u5199\u51fa,\u4f46\u9700\u8981\u9760\u63a8\u7406\u5f97\u51fa\u7684\u5b9e\u4f53<\/p>\n\n\n\n

\u8bba\u6587\u8981\u89e3\u51b3\u7684\u6838\u5fc3\u95ee\u9898\u53ef\u4ee5\u6d53\u7f29\u6210\u4e00\u53e5:
\u5982\u4f55\u8ba9 LLM \u5728\u4e0d\u4f9d\u8d56”\u8bad\u7ec3\u96c6\u4e2d\u6709\u540c\u683c\u5f0f\u6837\u672c”\u7684\u524d\u63d0\u4e0b,\u65e2\u80fd\u5229\u7528\u6587\u6863\u7248\u9762\u7ebf\u7d22\u3001\u53c8\u80fd\u4fdd\u7559\u63a8\u7406\u80fd\u529b,\u5b8c\u6210\u5bf9\u5f02\u6784 VRD \u7684\u9ad8\u8d28\u91cf\u4fe1\u606f\u62bd\u53d6(\u5305\u62ec\u9690\u542b\u503c\u63a8\u7406)\u3002
\u66f4\u5177\u4f53\u5730,\u4f5c\u8005\u628a\u5b83\u62c6\u6210\u51e0\u4e2a\u5b50\u95ee\u9898:
\u600e\u6837\u5728 LLM \u63d0\u793a\u91cc”\u88c5\u4e0b”\u8db3\u591f\u7684\u7248\u9762\u4e0e\u4e0a\u4e0b\u6587\u4fe1\u606f,\u800c\u4e0d\u8fc7\u8f7d?
\u600e\u6837\u8ba9\u6807\u6ce8\u8bad\u7ec3\u96c6\u4e2d\u7684\u63a8\u7406\u8fc7\u7a0b\u80fd\u8fc1\u79fb\u5230\u672a\u89c1\u683c\u5f0f?
\u600e\u6837\u8ba9\u5c0f\u6a21\u578b(7B\/14B)\u4e5f\u80fd\u591f\u8fbe\u5230\u5927\u6a21\u578b\u7684\u6c34\u5e73?
\u600e\u6837\u8ba9 LLM \u505a\u5230”\u6587\u6863\u91cc\u6ca1\u660e\u5199\u4f46\u80fd\u63a8\u51fa”\u7684\u5b9e\u4f53\u62bd\u53d6?<\/p>\n\n\n\n

\u89e3\u51b3\u529e\u6cd5- BLOCKIE<\/p>\n\n\n\n

\u5b9a\u4e49<\/strong>\uff1a\u4e0d\u53ef\u518d\u5206\u7684\u89c6\u89c9\u533a\u57df\u3002\u6587\u5b57\u5728\u7a7a\u95f4\u4e0a\u90bb\u8fd1\u4e14\u6c34\u5e73\/\u5782\u76f4\u5bf9\u9f50\uff0c\u62c6\u5f00\u5c31\u4f1a\u5931\u53bb\u542b\u4e49\u3002
\u4e3e\u4f8b<\/strong>\uff1a\u201cTOTAL ITEMS\u201d \u662f\u4e00\u4e2a\u539f\u5b50\uff0c\u62c6\u6210 \u201cTOTAL\u201d \u548c \u201cITEMS\u201d \u540e\u8bed\u4e49\u53d8\u6a21\u7cca\u3002
(2) Semantic Block\uff08\u8bed\u4e49\u5757\uff09\u2014\u2014 \u5168\u6587\u6700\u5173\u952e\u7684\u5b9a\u4e49
\u5f62\u5f0f\u5316\u5b9a\u4e49<\/strong>\uff1a\u5bf9\u6587\u6863 $D$ \u4e2d\u7684\u67d0\u4e2a\u7247\u6bb5 $B$\uff0c\u8bbe $v(B, C)$ \u8868\u793a\u5728\u4e0a\u4e0b\u6587 $C$ \u4e0b\u89e3\u6790 $B$ \u5f97\u5230\u7684 schema \u53d6\u503c\uff0c\u5219 $B$ \u662f\u8bed\u4e49\u5757\u5f53\u4e14\u4ec5\u5f53\uff1a
$v(B, B) = v(B, D) = V_{\\mathbb{E}}(B)$
\u901a\u4fd7\u7406\u89e3<\/strong>\uff1a\u8bed\u4e49\u5757\u5c31\u662f\u201c\u8131\u79bb\u5168\u6587\u4e5f\u80fd\u88ab\u72ec\u7acb\u6b63\u786e\u89e3\u6790\u201d\u7684\u6700\u5c0f\u81ea\u5305\u542b\u5355\u5143\u3002
\u4e3e\u4f8b<\/strong>\uff1a
\u662f\u8bed\u4e49\u5757<\/strong>\uff1a(SUB TOTAL 28.000)
\u4e0d\u662f\u8bed\u4e49\u5757<\/strong>\uff1a(COCONUT JELLY (L), 4.000) \u2014\u2014 \u8131\u79bb\u5b83\u6240\u5c5e\u7684\u4e3b\u83dc\u884c\uff0c\u65e0\u6cd5\u5224\u65ad\u5b83\u662f\u4e0d\u662f\u5b50\u9879\u3001\u5c5e\u4e8e\u54ea\u4e00\u884c\u3002
(3) Linkage\uff08\u539f\u5b50\u4e4b\u95f4\u7684\u4e24\u79cd\u94fe\u63a5\uff09
attribute:value \u94fe\u63a5<\/strong>\uff1a\u201cTOTAL ITEMS\u201d \u2194 \u201c1\u201d
hierarchical \u94fe\u63a5<\/strong>\uff1a\u4e3b\u83dc\u884c \u2194 \u5b83\u7684\u5b50\u9879
\u6838\u5fc3\u7ed3\u8bba<\/strong>\uff1a\u4e00\u4e2a\u8bed\u4e49\u5757 = \u4e00\u7ec4\u8bed\u4e49\u539f\u5b50\uff0c\u4e14\u7ec4\u5185\u6bcf\u4e2a\u539f\u5b50\u7684\u6240\u6709\u94fe\u63a5\u90fd\u5c01\u95ed\u5728\u7ec4\u5185\u3002<\/p>\n\n\n\n

3.2 \u4e09\u9636\u6bb5 Pipeline\uff08\u8bba\u6587\u7b2c 4 \u8282\uff09<\/h3>\n\n\n\n

\u6574\u4e2a\u6d41\u7a0b\u6a21\u62df\u4e86\u4eba\u9605\u8bfb\u6587\u6863\u7684\u65b9\u5f0f\uff1a\u5148\u770b\u5c40\u90e8 \u2192 \u518d\u62fc\u5168\u5c40\u3002<\/p>\n\n\n\n

\u9636\u6bb5 0\uff1aTrain Dataset Labeling\uff08\u79bb\u7ebf\u51c6\u5907\uff09 \u5229\u7528\u8bad\u7ec3\u96c6\u7684 key-value \u6807\u7b7e\uff0c\u8ba9 LLM\uff08\u8bba\u6587\u7528 Sonnet\uff09\u53cd\u5411\u751f\u6210\u4e09\u6837\u4e1c\u897f\uff1a<\/h4>\n\n\n\n
    \n
  • \u4e3a\u4ec0\u4e48\u628a\u8fd9\u6bb5\u6587\u5b57\u5224\u4e3a\u4e00\u4e2a\u5757\u7684 step-by-step reason\uff08\u9010\u6b65\u63a8\u7406\uff09<\/li>\n\n\n\n
  • \u5757\u5185\u7684\u6587\u672c<\/li>\n\n\n\n
  • \u5757\u5bf9\u5e94\u7684 partial annotation\uff08\u90e8\u5206\u586b\u597d\u7684 schema\uff09
    \u8fd9\u4e00\u6b65\u672c\u8d28\u662f\u628a\u201c\u6807\u7b7e + schema\u201d\u7ffb\u8bd1\u6210\u201c\u5757 + \u63a8\u7406\u8fc7\u7a0b\u201d\uff0c\u4f9b\u4e0b\u6e38\u505a few-shot\u3002<\/li>\n<\/ul>\n\n\n\n

    \u9636\u6bb5 1\uff1aBlock Creation\uff08\u5757\u521b\u5efa\uff09<\/h4>\n\n\n\n
      \n
    • \u8f93\u5165<\/strong>\uff1adocument schema + OCR text + bounding boxes + \u7528 OCR \u6587\u672c\u4f59\u5f26\u76f8\u4f3c\u5ea6\u9009\u51fa\u6765\u7684 5 \u4e2a few-shot \u5757\u793a\u4f8b<\/li>\n\n\n\n
    • \u4efb\u52a1<\/strong>\uff1a\u8ba9 LLM \u628a\u5f53\u524d\u6d4b\u8bd5\u6587\u6863\u5207\u6210\u82e5\u5e72\u8bed\u4e49\u5757\uff0c\u5e76\u8981\u6c42\u8f93\u51fa\u63a8\u7406<\/li>\n\n\n\n
    • \u5173\u952e<\/strong>\uff1a\u7528\u8bad\u7ec3\u96c6\u5757\u7684\u63a8\u7406\u8fc7\u7a0b\u201c\u6fc0\u53d1\u201d\u6d4b\u8bd5\u6587\u6863\u4e0a\u7c7b\u4f3c\u7684\u5206\u5757\u63a8\u7406<\/li>\n<\/ul>\n\n\n\n

      \u9636\u6bb5 2\uff1aBlock Parsing\uff08\u5757\u89e3\u6790\uff09<\/h4>\n\n\n\n
        \n
      • \u4efb\u52a1<\/strong>\uff1a\u5bf9\u6bcf\u4e2a\u5757\u72ec\u7acb\u89e3\u6790\uff0c\u8f93\u5165 schema + \u8be5\u5757\u6700\u76f8\u4f3c\u7684\u8bad\u7ec3\u96c6\u5757\u7684 few-shot<\/li>\n\n\n\n
      • \u4f18\u52bf<\/strong>\uff1a\u7531\u4e8e\u5757\u662f\u81ea\u5305\u542b\u7684\uff0c\u4e0d\u540c\u683c\u5f0f\u7684\u6587\u6863\u5e38\u5e38\u5171\u4eab\u76f8\u4f3c\u5757\uff08\u8bba\u6587 Figure 6\uff1a\u6cd5\u5f8b\u4e8b\u52a1\u6240\u4f20\u771f\u548c\u4fbf\u5229\u5e97\u53d1\u7968\u90fd\u6709\u201c\u8054\u7cfb\u65b9\u5f0f\u5757\u201d\uff09<\/li>\n\n\n\n
      • \u4f5c\u7528<\/strong>\uff1a\u8fd9\u91cc schema \u8d77\u5230\u201c\u7ed3\u6784\u5316\u8f93\u51fa\u683c\u5f0f\u7ea6\u675f\u201d\u7684\u4f5c\u7528<\/li>\n<\/ul>\n\n\n\n

        \u9636\u6bb5 3\uff1aBlock Combining\uff08\u5757\u7ec4\u5408\uff09<\/h4>\n\n\n\n
          \n
        • \u8f93\u5165<\/strong>\uff1a\u628a\u6240\u6709\u5757\u7684 partial parse + \u539f\u59cb OCR \u6587\u672c + bounding boxes + schema \u4e00\u8d77\u5582\u7ed9 LLM<\/li>\n\n\n\n
        • \u4efb\u52a1<\/strong>\uff1aLLM \u626e\u6f14 judge\uff0c\u5ba1\u89c6\u6bcf\u4e2a\u5757\u7684\u89e3\u6790\u63a8\u7406\uff0c\u628a\u788e\u7247\u62fc\u6210\u5b8c\u6574 schema \u8f93\u51fa<\/li>\n\n\n\n
        • \u5173\u952e<\/strong>\uff1a\u8fd9\u4e00\u6b65\u5904\u7406\u8de8\u5757\u4f9d\u8d56\uff08\u4f8b\u5982\u7ebf\u6027\u9879\u8ba1\u6570\uff09<\/li>\n<\/ul>\n\n\n\n

          3.3 \u63d0\u793a\u7b56\u7565<\/h3>\n\n\n\n

          \u6240\u6709\u63d0\u793a\u6a21\u677f\u57fa\u4e8e Claude 3.5 Sonnet \u8bbe\u8ba1\uff0c\u6545\u610f\u4e0d\u9488\u5bf9\u5176\u4ed6 LLM \u91cd\u65b0\u8c03\u53c2\uff0c\u4ee5\u8bc1\u660e BLOCKIE \u672c\u8eab\u7684\u63d0\u5347\u4e0d\u662f prompt tuning \u7684\u529f\u52b3\uff08\u9644\u5f55 A \u7ed9\u4e86\u4e09\u9636\u6bb5\u5b8c\u6574\u6a21\u677f\uff09\u3002<\/p>\n\n\n\n

          \u5b9e\u9a8c<\/p>\n\n\n\n

          \u6570\u636e\u96c6:CORD(\u5370\u5c3c\u9910\u996e\u6536\u636e,30 \u4e2a\u5c42\u7ea7\u5b9e\u4f53)\u3001FUNSD(\u8868\u5355,\u505a entity linking)\u3001SROIE(\u626b\u63cf\u6536\u636e)
          \u57fa\u6a21 LLM:Claude 3.5 Sonnet,Qwen 2.5 (7B\/14B\/32B\/72B)
          \u57fa\u7ebf:Layout-aware(LayoutLMv3, GeoLayoutLM, ESP, ERNIE-Layout, FormNetV2, RORE-GeoLayoutLM)+ LLM(DocLLM, LayoutLLM, LMDX-Gemini Pro, Sonnet \u96f6\/\u5c11\u6837\u672c)
          \u8bc4\u4ef7\u6307\u6807:Micro-F1
          Few-shot \u6570\u91cf:5<\/p>\n\n\n\n

          \"\"<\/figure>\n\n\n\n

          SOTA \u6027\u80fd\u9a8c\u8bc1(Table 1)<\/p>\n\n\n\n

          \"\"<\/figure>\n\n\n\n
          \"\"<\/figure>\n\n\n\n
          \"\"<\/figure>\n\n\n\n
          \"\"<\/figure>\n\n\n\n

          \u683c\u5f0f\u5f02\u6784\u4e0e\u672a\u89c1\u683c\u5f0f\u9c81\u68d2\u6027(Table 3)
          \u5757\u521b\u5efa\u662f\u5173\u952e\u74f6\u9888(Table 4 + Table 5)<\/p>\n\n\n\n

          \u5b9e\u9a8c 1 \u7acb SOTA \u2192 \u5b9e\u9a8c 2 \u6392\u9664”LLM \u5f3a=\u65b9\u6cd5\u5f3a”\u7684\u6df7\u6dc6 \u2192 \u5b9e\u9a8c 3 \u9a8c\u8bc1 desiderata \u4e2d\u7684”\u5f02\u6784+\u65b0\u683c\u5f0f”\u4e24\u6761 \u2192 \u5b9e\u9a8c 4 \u5b9a\u4f4d\u65b9\u6cd5\u5185\u7684\u5173\u952e\u6a21\u5757 \u2192 \u5b9e\u9a8c 5 \u9a8c\u8bc1 desiderata \u4e2d\u7684”value-absent inference”\u3002\u6574\u4e2a\u5b9e\u9a8c\u94fe\u6761\u7d27\u6263\u5f00\u7bc7\u63d0\u51fa\u7684\u56db\u6761 desiderata,\u903b\u8f91\u975e\u5e38\u4e25\u5bc6\u3002<\/p>\n\n\n\n

          \u6838\u5fc3\u5185\u5bb9\u603b\u7ed3\uff1a
          Semantic Block \u7684\u5f62\u5f0f\u5316\u5b9a\u4e49 \u2014\u2014 \u7b2c\u4e00\u6b21\u7ed9 VRD \u62bd\u53d6\u91cc\u7684”\u81ea\u5305\u542b\u5355\u5143”\u4e00\u4e2a\u6570\u5b66\u5b9a\u4e49 v(B,B)=v(B,D)
          ,\u628a”\u5757”\u4ece\u76f4\u89c9\u6982\u5ff5\u53d8\u6210\u53ef\u9a8c\u8bc1\u6982\u5ff5
          \u4ece\u5168\u5c40\u63a8\u7406\u5230\u5c40\u90e8\u63a8\u7406\u7684\u8303\u5f0f\u8f6c\u79fb \u2014\u2014 \u4e0d\u518d\u8ba9 LLM \u4e00\u6b21\u6027\u7406\u89e3\u6574\u5f20\u6587\u6863,\u800c\u662f\u5206\u5757\u72ec\u7acb\u63a8\u7406 + \u5168\u5c40\u7ec4\u5408,\u663e\u8457\u964d\u4f4e\u5355\u6b65\u4efb\u52a1\u590d\u6742\u5ea6
          \u57fa\u4e8e\u63a8\u7406\u7684\u8bad\u7ec3\u6837\u672c\u7ec4\u7ec7 \u2014\u2014 \u7528 LLM \u628a\u8bad\u7ec3\u96c6\u6807\u7b7e\u53cd\u5411\u751f\u6210”\u5757+step-by-step reason”,\u8fd9\u4e9b reason \u5728\u6d4b\u8bd5\u65f6\u901a\u8fc7 few-shot \u8fc1\u79fb,\u7b49\u4e8e\u628a”\u5982\u4f55\u5206\u5757\/\u5982\u4f55\u89e3\u6790”\u7684\u5143\u77e5\u8bc6\u4e5f\u53d8\u6210\u4e86\u53ef\u5b66\u4e60\u4fe1\u53f7
          \u8de8\u683c\u5f0f\u7684\u5757\u7ea7\u76f8\u4f3c\u68c0\u7d22 \u2014\u2014 \u4e0d\u540c\u6a21\u677f\u7684\u6587\u6863\u5176\u5b9e\u5171\u4eab\u76f8\u4f3c\u8bed\u4e49\u5757(\u4f20\u771f\u548c\u53d1\u7968\u90fd\u6709”\u8054\u7cfb\u4fe1\u606f\u5757”),\u8fd9\u8ba9 few-shot \u68c0\u7d22\u4e0d\u518d\u53d7\u9650\u4e8e”\u5fc5\u987b\u6709\u540c\u683c\u5f0f\u6837\u672c”
          \u5757\u521b\u5efa\u74f6\u9888\u7684\u5b9e\u8bc1\u5b9a\u4f4d \u2014\u2014 \u901a\u8fc7 GT-block \u6d88\u878d\u5b9e\u9a8c,\u9996\u6b21\u660e\u786e\u6307\u51fa”\u5757\u5207\u5bf9\u4e86,\u5c0f\u6a21\u578b\u4e5f\u884c;\u5757\u5207\u9519\u4e86,\u540e\u9762\u6551\u4e0d\u56de\u6765”<\/p>\n","protected":false},"excerpt":{"rendered":"

          \u4f5c\u8005\uff1aAniket Bhattacharyya1, Anurag Tripathi, Ujjal Das, A […]<\/p>\n","protected":false},"author":66,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-5174","post","type-post","status-publish","format-standard","hentry","category-events"],"_links":{"self":[{"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/posts\/5174","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/users\/66"}],"replies":[{"embeddable":true,"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/comments?post=5174"}],"version-history":[{"count":1,"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/posts\/5174\/revisions"}],"predecessor-version":[{"id":5180,"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/posts\/5174\/revisions\/5180"}],"wp:attachment":[{"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/media?parent=5174"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/categories?post=5174"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.15zhi.net\/blog\/wp-json\/wp\/v2\/tags?post=5174"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}