{"id":3243257,"date":"2026-08-22T18:45:31","date_gmt":"2026-08-22T10:45:31","guid":{"rendered":"http:\/\/www.nlpir.org\/wordpress\/?p=3243257"},"modified":"2026-08-22T18:45:31","modified_gmt":"2026-08-22T10:45:31","slug":"nlpir%e5%ae%9e%e9%aa%8c%e5%ae%a4%e5%a4%9a%e7%af%87%e8%ae%ba%e6%96%87%e8%a2%abemnlp2026%e5%bd%95%e7%94%a8","status":"publish","type":"post","link":"http:\/\/www.nlpir.org\/wordpress\/2026\/08\/22\/nlpir%e5%ae%9e%e9%aa%8c%e5%ae%a4%e5%a4%9a%e7%af%87%e8%ae%ba%e6%96%87%e8%a2%abemnlp2026%e5%bd%95%e7%94%a8\/","title":{"rendered":"NLPIR\u5b9e\u9a8c\u5ba4\u591a\u7bc7\u8bba\u6587\u88abEMNLP2026\u5f55\u7528"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>\u5feb\u8baf<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u8fd1\u65e5\uff0c\u81ea\u7136\u8bed\u8a00\u5904\u7406\u9886\u57df\u56fd\u9645\u9876\u7ea7\u4f1a\u8bae Conference on Empirical Methods in Natural Language Processing\uff08EMNLP\uff09\u516c\u5e03\u8bba\u6587\u5f55\u7528\u7ed3\u679c\uff0cNLPIR\u5b9e\u9a8c\u5ba4\u5171\u67094\u7bc7\u8bba\u6587\u88ab\u63a5\u6536\u3002\u76f8\u5173\u7814\u7a76\u8986\u76d6\u5927\u8bed\u8a00\u6a21\u578b\u63a8\u7406\u3001\u97f3\u9891\u8bed\u8a00\u6a21\u578b\u3001\u591a\u6a21\u6001\u7406\u89e3\u3001\u65f6\u95f4\u8ba4\u77e5\u63a8\u7406\u3001\u5f00\u653e\u5173\u7cfb\u62bd\u53d6\u7b49\u65b9\u5411\uff0c\u5c55\u73b0\u4e86\u5b9e\u9a8c\u5ba4\u5728\u8ba4\u77e5\u667a\u80fd\u3001\u5927\u6a21\u578b\u4e0e\u591a\u6a21\u6001\u4eba\u5de5\u667a\u80fd\u9886\u57df\u7684\u6301\u7eed\u63a2\u7d22\u4e0e\u521b\u65b0\u80fd\u529b\u3002\u4ee5\u4e0b\u4e3a\u5165\u9009\u8bba\u6587\u4ecb\u7ecd\u3002<\/p>\n\n\n\n<div class=\"wp-block-group is-nowrap is-layout-flex wp-container-core-group-is-layout-8f761849 wp-block-group-is-layout-flex\">\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4e00<\/strong><\/p>\n<\/div>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u9898\u76ee\uff1a<\/strong>Is the Future Predictable? Modeling Future Predictability for Reliable Temporal Reasoning<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4f5c\u8005\uff1a<\/strong>\u9ec4\u548f\u4eea \u5f20\u534e\u5e73 \u5d14\u6587\u8000 \u674e\u79cb\u6c60 \u674e\u78ca \u5f20\u5b9d\u534e<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u4f5c\u8005\u5355\u4f4d\uff1a<\/strong>\u5317\u4eac\u7406\u5de5\u5927\u5b66<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u6982\u8ff0\uff1a<\/strong>\u5927\u8bed\u8a00\u6a21\u578b\uff08LLMs\uff09\u5728\u8fdb\u884c\u672a\u6765\u63a8\u7406\u65f6\u4ecd\u7136\u5b58\u5728\u53ef\u9760\u6027\u4e0d\u8db3\u7684\u95ee\u9898\u3002\u73b0\u6709\u6a21\u578b\u901a\u5e38\u5ffd\u89c6\u4e0d\u540c\u672a\u6765\u72b6\u6001\u4e4b\u95f4\u7684\u5dee\u5f02\uff0c\u5373\u4f7f\u6240\u67e5\u8be2\u7684\u4e8b\u5b9e\u672c\u8eab\u5177\u6709\u9ad8\u5ea6\u4e0d\u786e\u5b9a\u6027\uff0c\u6a21\u578b\u4ecd\u7136\u4f1a\u7ed9\u51fa\u786e\u5b9a\u6027\u7684\u7b54\u6848\u3002\u4e3a\u5e94\u5bf9\u8fd9\u4e00\u6311\u6218\uff0c\u672c\u6587\u63d0\u51fa\u4e86 ChronoRegime\uff0c\u4e00\u79cd\u9762\u5411\u672a\u6765\u7684\u65f6\u95f4\u63a8\u7406\u4efb\u52a1\uff0c\u8981\u6c42\u6a21\u578b\u5224\u65ad\uff1a\u5df2\u6709\u7684\u5386\u53f2\u65f6\u95f4\u8bc1\u636e\u662f\u5426\u8db3\u4ee5\u652f\u6301\u5bf9\u672a\u6765\u7b54\u6848\u505a\u51fa\u786e\u5b9a\u6027\u5224\u65ad\u3002\u6b64\u5916\uff0c\u672c\u6587\u63d0\u51fa\u4e86 \u7f6e\u4fe1\u611f\u77e5\u65f6\u95f4\u63a8\u7406\u6846\u67b6\uff08Confidence-Aware Temporal Reasoning, CTR\uff09\u3002\u8be5\u6846\u67b6\u901a\u8fc7\u5206\u6790\u65f6\u95f4\u6f14\u5316\u89c4\u5f8b\u6765\u8bc4\u4f30\u672a\u6765\u72b6\u6001\u7684\u53ef\u9884\u6d4b\u6027\uff0c\u5e76\u636e\u6b64\u8c03\u8282\u6a21\u578b\u662f\u5426\u5e94\u8be5\u5bf9\u7b54\u6848\u505a\u51fa\u627f\u8bfa\u3002\u4e0e\u65e0\u6761\u4ef6\u5730\u8fdb\u884c\u672a\u6765\u9884\u6d4b\u4e0d\u540c\uff0cCTR\u4ec5\u5728\u91cd\u6784\u540e\u7684\u8bc1\u636e\u80fd\u591f\u652f\u6301\u76ee\u6807\u672a\u6765\u72b6\u6001\u65f6\u624d\u7ed9\u51fa\u786e\u5b9a\u6027\u9884\u6d4b\u3002\u5728\u591a\u4e2a\u5927\u8bed\u8a00\u6a21\u578b\u4e0a\u7684\u5b9e\u9a8c\u8868\u660e\uff0c\u4e0e\u666e\u901a\u95ee\u7b54\u57fa\u7ebf\u65b9\u6cd5\u76f8\u6bd4\uff0cCTR\u4f7f\u4e25\u683c\u603b\u4f53\u51c6\u786e\u7387\u5e73\u5747\u63d0\u5347 6.7\u4e2a\u767e\u5206\u70b9\uff0c\u5e76\u4f7f\u5e7b\u89c9\u7387\u5e73\u5747\u964d\u4f4e 18.4\u4e2a\u767e\u5206\u70b9\u3002\u5b9e\u9a8c\u7ed3\u679c\u8bc1\u660e\uff0c\u663e\u5f0f\u8bc4\u4f30\u672a\u6765\u4e8b\u4ef6\u7684\u53ef\u9884\u6d4b\u6027\uff0c\u662f\u63d0\u5347\u672a\u6765\u4e0d\u786e\u5b9a\u73af\u5883\u4e0b\u65f6\u95f4\u63a8\u7406\u5b89\u5168\u6027\u548c\u53ef\u9760\u6027\u7684\u6709\u6548\u673a\u5236\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/yozlYZeLGPfic1EazL1oqCvFxib6UpUicRNW6ymL0u107JNuibtSa9wM0oA2xicAy0PDbta29yibJXQefvPpTXbs5mUcSJgWtkCJX50U4PO5gWdU8\/640?wx_fmt=png&amp;from=appmsg&amp;tp=wxpic&amp;wxfrom=5&amp;wx_lazy=1#imgIndex=3\" alt=\"\u56fe\u7247\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4e8c<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u9898\u76ee\uff1a<\/strong>Learning to Explore the Crux in Formal Theorem Proving<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4f5c\u8005\uff1a<\/strong>\u674e\u6797\u7ff0 \u5d14\u6587\u8000 \u5f20\u516c\u94ce \u987e\u7acb\u5b8f \u5468\u7433 \u5546\u5efa\u4e91 \u5f20\u534e\u5e73<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u4f5c\u8005\u5355\u4f4d\uff1a<\/strong>\u5317\u4eac\u7406\u5de5\u5927\u5b66\u3001\u8682\u8681\u96c6\u56e2<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u6982\u8ff0\uff1a<\/strong>\u4f5c\u8005\u63d0\u51fa\u4e86\u4e00\u79cd\u9762\u5411\u5f62\u5f0f\u5316\u5b9a\u7406\u8bc1\u660e\u7684\u57fa\u4e8e\u71b5\u7684\u6700\u4f73\u4f18\u5148\u641c\u7d22\u7b97\u6cd5\uff08Entropy-Based Best-First Search, EBFS\uff09\uff0c\u9488\u5bf9Lean\u5f62\u5f0f\u5316\u8bc1\u660e\u8fc7\u7a0b\u4e2d\u4e0d\u540c\u63a8\u7406\u6b65\u9aa4\u96be\u5ea6\u548c\u91cd\u8981\u6027\u4e0d\u5747\u8861\u7684\u95ee\u9898\uff0c\u4ece\u71b5\u7684\u89c6\u89d2\u8bc6\u522b\u8bc1\u660e\u6811\u4e2d\u66f4\u503c\u5f97\u63a2\u7d22\u7684\u5173\u952e\u6b65\u9aa4\uff0c\u5e76\u636e\u6b64\u4f18\u5316\u8bc1\u660e\u641c\u7d22\u8fc7\u7a0b\uff0c\u4f7f\u8bc1\u660e\u751f\u6210\u6a21\u578b\u80fd\u591f\u5728\u6709\u9650\u641c\u7d22\u9884\u7b97\u4e0b\u66f4\u9ad8\u6548\u5730\u63a2\u7d22\u8bc1\u660e\u7a7a\u95f4\uff0c\u63d0\u9ad8\u627e\u5230\u6b63\u786e\u8bc1\u660e\u7684\u6982\u7387\u3002\u5728\u6b64\u57fa\u7840\u4e0a\uff0c\u4f5c\u8005\u8fdb\u4e00\u6b65\u63d0\u51fa\u4e86\u9762\u5411\u8bc1\u660e\u6811\u641c\u7d22\u7684\u5f3a\u5316\u5fae\u8c03\u65b9\u6cd5\uff0c\u5e76\u5728\u751f\u6210\u9636\u6bb5\u7ed3\u5408EBFS\u8fdb\u4e00\u6b65\u63d0\u5347\u6a21\u578b\u7684\u5f62\u5f0f\u5316\u8bc1\u660e\u80fd\u529b\u3002\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\uff0c\u5728\u76f8\u540c\u91c7\u6837\u9884\u7b97\u4e0b\uff0c\u8be5\u65b9\u6cd5\u76f8\u8f83\u4e8e\u5b8c\u6574\u8bc1\u660e\u751f\u6210\u57fa\u7ebf\uff0c\u5728ProofNet\u6d4b\u8bd5\u96c6\u4e0a\u63d0\u53473.2%\uff0c\u5e76\u5728MiniF2F\u548cProverBench\u4e0a\u5206\u522b\u63d0\u53471.9%\u548c1.5%\uff0c\u9a8c\u8bc1\u4e86\u8be5\u65b9\u6cd5\u5728\u5f62\u5f0f\u5316\u5b9a\u7406\u8bc1\u660e\u641c\u7d22\u6548\u7387\u548c\u8bc1\u660e\u6210\u529f\u7387\u65b9\u9762\u7684\u6709\u6548\u6027\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"https:\/\/mmbiz.qpic.cn\/sz_mmbiz_png\/yozlYZeLGPc7wx5mq3CRhlAxO7z2u2tLvAVyJribwMiacWbDYJmsuZnfo7Udx2v3AjmAT4OqHxeRgJCqlHGfSCZicSwiasNicK1K7BQ88HbTP55c\/640?wx_fmt=png&amp;from=appmsg&amp;tp=wxpic&amp;wxfrom=5&amp;wx_lazy=1#imgIndex=5\" alt=\"\u56fe\u7247\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4e09<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u9898\u76ee<\/strong>\uff1aModality-Specific Gating for Large Audio-Language Models<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4f5c\u8005\uff1a<\/strong>\u4efb\u73cd\u59ae \u738b\u5a1f \u674e\u79cb\u6c60 \u674e\u78ca \u9ad8\u6625\u6653 \u5f20\u5b9d\u534e \u5f20\u534e\u5e73<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u4f5c\u8005\u5355\u4f4d\uff1a<\/strong>\u5317\u4eac\u7406\u5de5\u5927\u5b66<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u6982\u8ff0\uff1a<\/strong>\u5927\u578b\u97f3\u9891\u8bed\u8a00\u6a21\u578b\uff08Large Audio-Language Models, LALMs\uff09\u8fd1\u5e74\u6765\u5728\u97f3\u9891-\u6587\u672c\u6307\u4ee4\u7406\u89e3\u548c\u591a\u6a21\u6001\u4ea4\u4e92\u65b9\u9762\u53d6\u5f97\u4e86\u663e\u8457\u8fdb\u5c55\u3002\u7136\u800c\uff0c\u7531\u4e8e\u5176\u91c7\u7528\u6587\u672c\u6761\u4ef6\u751f\u6210\u67b6\u6784\uff0c\u5f53\u524d\u7684LALM\u5f80\u5f80\u4f1a\u53d7\u5230\u8f93\u5165\u6587\u672c\u63d0\u793a\u7684\u8fc7\u5ea6\u5f71\u54cd\u3002\u5f53\u58f0\u5b66\u8bc1\u636e\u4e0e\u6587\u672c\u63d0\u793a\u53d1\u751f\u51b2\u7a81\u65f6\uff0c\u9519\u8bef\u6216\u65e0\u5173\u7684\u6587\u672c\u4fe1\u606f\u53ef\u80fd\u4f1a\u5f15\u5bfc\u6a21\u578b\u504f\u79bb\u771f\u5b9e\u7684\u97f3\u9891\u5185\u5bb9\uff0c\u5bfc\u81f4\u9884\u6d4b\u7ed3\u679c\u9519\u8bef\u3002\u672c\u6587\u7cfb\u7edf\u5206\u6790\u4e86\u8fd9\u4e00\u95ee\u9898\uff0c\u5e76\u53d1\u73b0\u5177\u6709\u4ee3\u8868\u6027\u7684LALM\u5728\u751f\u6210\u8fc7\u7a0b\u4e2d\u5b58\u5728\u6301\u7eed\u6027\u7684\u97f3\u9891-\u6587\u672c\u6a21\u6001\u5931\u8861\u73b0\u8c61\uff1a\u968f\u7740\u751f\u6210\u6b65\u9aa4\u63a8\u8fdb\uff0c\u6a21\u578b\u5bf9\u97f3\u9891token\u7684\u5173\u6ce8\u7a0b\u5ea6\u9010\u6e10\u51cf\u5f31\u3002\u4e3a\u89e3\u51b3\u8fd9\u4e00\u5931\u8861\u95ee\u9898\uff0c\u672c\u6587\u63d0\u51fa\u4e86\u6a21\u6001\u7279\u5b9a\u95e8\u63a7\u673a\u5236\uff08Modality-Specific Gating, MSG\uff09\u3002\u8be5\u65b9\u6cd5\u662f\u4e00\u79cd\u58f0\u5b66\u6821\u51c6\u673a\u5236\uff0c\u4ec5\u5f15\u5165\u5c11\u91cf\u989d\u5916\u53c2\u6570\uff0c\u5373\u53ef\u589e\u5f3a\u6a21\u578b\u5bf9\u97f3\u9891\u4fe1\u606f\u7684\u611f\u77e5\u80fd\u529b\u3002\u5177\u4f53\u800c\u8a00\uff0cMSG\u5728softmax\u8ba1\u7b97\u4e4b\u524d\uff0c\u9488\u5bf9\u97f3\u9891token\u7684\u6ce8\u610f\u529blogit\u5f15\u5165\u4e00\u79cd\u57fa\u4e8e\u67e5\u8be2\u6761\u4ef6\uff08query-conditioned\uff09\u7684\u9010token\u504f\u7f6e\uff08token-wise bias\uff09\uff0c\u4ece\u800c\u63d0\u5347\u97f3\u9891\u4fe1\u606f\u5728\u6ce8\u610f\u529b\u8ba1\u7b97\u4e2d\u7684\u6743\u91cd\uff0c\u540c\u65f6\u4fdd\u6301\u539f\u6709\u6ce8\u610f\u529b\u8ba1\u7b97\u8fc7\u7a0b\u4e0d\u53d8\u3002\u901a\u8fc7\u53c2\u6570\u9ad8\u6548\u5fae\u8c03\uff08parameter-efficient fine-tuning\uff09\uff0cMSG\u80fd\u591f\u589e\u5f3a\u6a21\u578b\u5bf9\u58f0\u5b66\u8bc1\u636e\u7684\u5229\u7528\u80fd\u529b\uff0c\u4f7f\u5176\u5728\u58f0\u97f3\u5206\u7c7b\uff08vocal sound classification\uff09\u548c\u8bed\u97f3\u60c5\u611f\u8bc6\u522b\uff08speech emotion recognition\uff09\u4efb\u52a1\u4e2d\u6301\u7eed\u83b7\u5f97\u6027\u80fd\u63d0\u5347\uff0c\u5e76\u4e14\u5728\u9762\u5bf9\u5bf9\u6297\u6027\u6587\u672c\u63d0\u793a\u6216\u65e0\u5173\u6587\u672c\u63d0\u793a\u65f6\u8868\u73b0\u51fa\u66f4\u5f3a\u7684\u9c81\u68d2\u6027\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/yozlYZeLGPfSNfQyiaAacaFic2Mwic4vPpOs4RxicNia4jnPDYiaWicg24oYict1a0hJSS3yasDq7rgClZeGvIa3iaf1FJyRA4c6c65L4A7cTIGq2aW4\/640?wx_fmt=png&amp;from=appmsg&amp;tp=wxpic&amp;wxfrom=5&amp;wx_lazy=1#imgIndex=7\" alt=\"\u56fe\u7247\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u56db<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u9898\u76ee\uff1a<\/strong>Distributional Recalibration for Zero-Shot Relation Extraction<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u4f5c\u8005\uff1a<\/strong>\u5218\u6c38\u946b \u5f20\u534e\u5e73 \u674e\u79cb\u6c60 \u674e\u78ca \u9ad8\u6625\u6653 \u5415\u6d69\u6210 \u4e25\u82e5\u8c6a \u5f20\u5b9d\u534e<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u4f5c\u8005\u5355\u4f4d\uff1a<\/strong>\u5317\u4eac\u7406\u5de5\u5927\u5b66 \u65b0\u7586\u5927\u5b66<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u8bba\u6587\u6982\u8ff0\uff1a<\/strong>\u57fa\u4e8e\u8054\u5408\u7f16\u7801\uff08joint-encoding\uff09\u7684\u96f6\u6837\u672c\u5173\u7cfb\u62bd\u53d6\u65b9\u6cd5\u80fd\u591f\u5728\u4e00\u6b21Transformer\u524d\u5411\u4f20\u64ad\u4e2d\u5bf9\u6240\u6709\u5019\u9009\u5173\u7cfb\u6807\u7b7e\u8fdb\u884c\u5206\u7c7b\uff0c\u56e0\u6b64\u517c\u5177\u8f83\u9ad8\u7684\u51c6\u786e\u7387\u548c\u63a8\u7406\u6548\u7387\u3002\u7136\u800c\uff0c\u5728\u5b8c\u6574\u5173\u7cfb\u8bcd\u8868\u89c4\u6a21\uff08full-vocabulary scale\uff09\u4e0b\uff0c\u8be5\u65b9\u6cd5\u4f1a\u66b4\u9732\u51fa\u4e24\u7c7b\u5206\u5e03\u6027\u95ee\u9898\uff1a\u4e00\u65b9\u9762\uff0c\u672a\u5f52\u4e00\u5316\u7684\u70b9\u79ef\u8bc4\u5206\uff08unnormalized dot-product scoring\uff09\u4f1a\u5bfc\u81f4\u53ec\u56de\u574d\u7f29\uff08recall collapse\uff09\uff0c\u4f7f\u90e8\u5206\u5173\u7cfb\u7c7b\u578b\u65e0\u6cd5\u88ab\u6a21\u578b\u9884\u6d4b\uff1b\u53e6\u4e00\u65b9\u9762\uff0c\u5b9e\u4f8b\u7ea7\u4e8c\u5143\u4ea4\u53c9\u71b5\u635f\u5931\uff08per-instance BCE loss\uff09\u4f1a\u5bfc\u81f4\u771f\u5b9e\u7c7b\u522b\u4e0e\u8d1f\u7c7b\u522b\u8bc4\u5206\u5206\u5e03\u4e25\u91cd\u6df7\u53e0\uff0c\u96be\u4ee5\u786e\u5b9a\u53ef\u9760\u9608\u503c\uff0c\u9020\u6210\u7f6e\u4fe1\u5ea6\u6b67\u4e49\uff08confidence ambiguity\uff09\u3002\u9488\u5bf9\u4e0a\u8ff0\u95ee\u9898\uff0c\u672c\u6587\u63d0\u51fa\u4e86SCAR\uff08Stabilized Cascade and Alignment for Robust joint-encoding ZSRE\uff09\u65b9\u6cd5\u3002\u8be5\u65b9\u6cd5\u4fdd\u6301\u539f\u6709\u57fa\u7ebf\u6a21\u578b\u7ed3\u6784\u4e0d\u53d8\uff0c\u901a\u8fc7\u4e24\u4e2a\u7ecf\u8fc7\u7a33\u5b9a\u5316\u8bbe\u8ba1\u7684\u8f85\u52a9\u5206\u652f\u5bf9\u8bc4\u5206\u5206\u5e03\u548c\u8868\u793a\u5206\u5e03\u8fdb\u884c\u91cd\u65b0\u6821\u51c6\uff1a\u4e00\u662f\u57fa\u4e8e\u68af\u5ea6\u9694\u79bb\u7684\u7c97\u5230\u7ec6\u7ea7\u8054\u6a21\u5757\uff0c\u901a\u8fc7Z-score\u5f52\u4e00\u5316\u8c03\u6574\u5173\u7cfb\u8bc4\u5206\u5206\u5e03\uff0c\u7f13\u89e3\u53ec\u56de\u574d\u7f29\u95ee\u9898\uff1b\u4e8c\u662f\u4ec5\u7528\u4e8e\u8bad\u7ec3\u9636\u6bb5\u7684\u5bf9\u6bd4\u5bf9\u9f50\u6a21\u5757\uff0c\u901a\u8fc7\u4f18\u5316\u5173\u7cfb\u8868\u793a\u7a7a\u95f4\uff0c\u63d0\u9ad8\u4e0d\u540c\u5173\u7cfb\u7c7b\u522b\u4e4b\u95f4\u7684\u533a\u5206\u80fd\u529b\u3002\u5728Wiki-ZSL\u6570\u636e\u96c6\uff08\u5019\u9009\u5173\u7cfb\u6570\u91cf m=15\uff09\u4e0a\u7684\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\uff0cSCAR\u5c06Macro F1\u6307\u6807\u4ece75.9%\u63d0\u5347\u81f383.0%\uff0c\u5e76\u5c06\u96f6\u53ec\u56de\u5173\u7cfb\u7c7b\u578b\u6570\u91cf\u4ece35\u7c7b\u964d\u4f4e\u81f31\u7c7b\u3002\u5b9e\u9a8c\u7ed3\u679c\u8bf4\u660e\uff0c\u4f20\u7edf\u7684\u5c0f\u89c4\u6a21\u5019\u9009\u5173\u7cfb\u8bc4\u6d4b\u534f\u8bae\u53ef\u80fd\u4f1a\u63a9\u76d6\u5927\u89c4\u6a21\u5173\u7cfb\u8bcd\u8868\u90e8\u7f72\u4e2d\u7684\u5173\u952e\u5931\u6548\u6a21\u5f0f\uff0c\u800cSCAR\u80fd\u591f\u6709\u6548\u63d0\u5347\u96f6\u6837\u672c\u5173\u7cfb\u62bd\u53d6\u65b9\u6cd5\u5728\u5f00\u653e\u73af\u5883\u4e0b\u7684\u7a33\u5b9a\u6027\u548c\u53ef\u9760\u6027\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"https:\/\/mmbiz.qpic.cn\/sz_mmbiz_png\/yozlYZeLGPeTqUlibSvd5eU7dtgAaTTcYY0eicW3m8galqPqvdmb1Kiaic19iayo8S8qCsfUXyn19sKfwG2u0evT5ichWcvNTxtj0aRfQUPNuUYSU\/640?wx_fmt=png&amp;from=appmsg&amp;tp=wxpic&amp;wxfrom=5&amp;wx_lazy=1#imgIndex=9\" alt=\"\u56fe\u7247\"\/><\/figure>\n","protected":false},"excerpt":{"rendered":"<p>\u5feb\u8baf \u8fd1\u65e5\uff0c\u81ea\u7136\u8bed\u8a00\u5904\u7406\u9886\u57df\u56fd\u9645\u9876\u7ea7\u4f1a\u8bae Conference on Empir &hellip; <a href=\"http:\/\/www.nlpir.org\/wordpress\/2026\/08\/22\/nlpir%e5%ae%9e%e9%aa%8c%e5%ae%a4%e5%a4%9a%e7%af%87%e8%ae%ba%e6%96%87%e8%a2%abemnlp2026%e5%bd%95%e7%94%a8\/\">\u7ee7\u7eed\u9605\u8bfb <span class=\"meta-nav\">&rarr;<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_elementor_edit_mode":"","_elementor_template_type":"","_elementor_data":"","_elementor_page_settings":null},"categories":[31,38],"tags":[],"class_list":["post-3243257","post","type-post","status-publish","format-standard","hentry","category-paper","category-38"],"_links":{"self":[{"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/posts\/3243257","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/comments?post=3243257"}],"version-history":[{"count":1,"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/posts\/3243257\/revisions"}],"predecessor-version":[{"id":3243267,"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/posts\/3243257\/revisions\/3243267"}],"wp:attachment":[{"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/media?parent=3243257"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/categories?post=3243257"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.nlpir.org\/wordpress\/wp-json\/wp\/v2\/tags?post=3243257"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}