{"id":21855,"date":"2024-08-23T10:33:42","date_gmt":"2024-08-23T02:33:42","guid":{"rendered":"https:\/\/aif.amtbbs.org\/?p=21855"},"modified":"2024-08-23T10:33:42","modified_gmt":"2024-08-23T02:33:42","slug":"%e7%94%a8ai%e8%87%aa%e5%8a%a8%e8%ae%be%e8%ae%a1%e6%99%ba%e8%83%bd%e4%bd%93%ef%bc%8c%e6%95%b0%e5%ad%a6%e6%8f%90%e5%88%8625-9%ef%bc%8c%e8%bf%9c%e8%b6%85%e6%89%8b%e5%b7%a5%e8%ae%be%e8%ae%a1","status":"publish","type":"post","link":"https:\/\/aif.amtbbs.org\/index.php\/2024\/08\/23\/21855\/","title":{"rendered":"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1"},"content":{"rendered":"<div><img data-dominant-color=\"c6bfbe\" data-has-transparency=\"false\" style=\"--dominant-color: #c6bfbe;\" loading=\"lazy\" decoding=\"async\" class=\"not-transparent alignnone size-full wp-image-21857\" src=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/08\/f161e17701b87070c8a11410758b765adcc949-300x167-1.jpg\" width=\"300\" height=\"167\" alt=\"\" srcset=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/08\/f161e17701b87070c8a11410758b765adcc949-300x167-1.jpg 300w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/08\/f161e17701b87070c8a11410758b765adcc949-300x167-1-150x84.jpg 150w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/div>\n<div class=\"article-desc\">\u57fa\u4e8e ADAS \u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u7684\u6027\u80fd\u5927\u5927\u4f18\u4e8e\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u57fa\u7ebf\u3002<\/div>\n<div id=\"postspictures\" class=\"article-content\">\n<div id=\"container\" class=\"container am-engine\" data-v-1d7a5742=\"\" data-element=\"root\">\n<p>\u57fa\u7840\u6a21\u578b (FM) \u5982 GPT \u548c Claude \uff0c\u6b63\u5728\u6210\u4e3a\u901a\u7528\u667a\u80fd\u4f53\u7684\u5f3a\u6709\u529b\u652f\u6301\uff0c\u88ab\u8d8a\u6765\u8d8a\u591a\u7684\u7528\u4e8e\u591a\u79cd\u63a8\u7406\u548c\u89c4\u5212\u4efb\u52a1\u3002<\/p>\n<p>\u7136\u800c\uff0c\u5728\u89e3\u51b3\u95ee\u9898\u65f6\uff0c\u9700\u8981\u7684\u667a\u80fd\u4f53\u901a\u5e38\u662f\u5177\u6709\u591a\u4e2a\u7ec4\u4ef6\u7684\u590d\u5408\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u800c\u4e0d\u662f\u5355\u7247\u6a21\u578b\u67e5\u8be2 \u3002\u6b64\u5916\uff0c\u4e3a\u4e86\u4f7f\u667a\u80fd\u4f53\u80fd\u591f\u89e3\u51b3\u590d\u6742\u7684\u73b0\u5b9e\u4e16\u754c\u4efb\u52a1\uff0c\u5b83\u4eec\u901a\u5e38\u9700\u8981\u8bbf\u95ee\u5916\u90e8\u5de5\u5177\uff0c\u4f8b\u5982\u641c\u7d22\u5f15\u64ce\u3001\u4ee3\u7801\u6267\u884c\u548c\u6570\u636e\u5e93\u67e5\u8be2\u3002<\/p>\n<p>\u56e0\u6b64\uff0c\u4eba\u4eec\u63d0\u51fa\u4e86\u8bb8\u591a\u6709\u6548\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\u6784\u5efa\u5757\uff0c\u4f8b\u5982\u601d\u7ef4\u94fe\u89c4\u5212\u548c\u63a8\u7406\u3001\u8bb0\u5fc6\u7ed3\u6784\u3001\u5de5\u5177\u4f7f\u7528\u548c\u81ea\u6211\u53cd\u601d\u3002\u5c3d\u7ba1\u8fd9\u4e9b\u667a\u80fd\u4f53\u5df2\u7ecf\u5728\u5404\u79cd\u5e94\u7528\u4e2d\u53d6\u5f97\u4e86\u663e\u8457\u7684\u6210\u529f\uff0c\u4f46\u5f00\u53d1\u8fd9\u4e9b\u6784\u5efa\u5757\u5e76\u5c06\u5b83\u4eec\u7ec4\u5408\u6210\u590d\u6742\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\u901a\u5e38\u9700\u8981\u7279\u5b9a\u9886\u57df\u7684\u624b\u52a8\u8c03\u6574\u4ee5\u53ca\u7814\u7a76\u4eba\u5458\u548c\u5de5\u7a0b\u5e08\u7684\u5927\u91cf\u52aa\u529b\u3002<\/p>\n<p>\u7136\u800c\uff0c\u673a\u5668\u5b66\u4e60\u7684\u5386\u53f2\u544a\u8bc9\u6211\u4eec\uff0c\u624b\u5de5\u8bbe\u8ba1\u7684\u89e3\u51b3\u65b9\u6848\u6700\u7ec8\u4f1a\u88ab\u6a21\u578b\u5b66\u4e60\u5230\u7684\u89e3\u51b3\u65b9\u6848\u6240\u53d6\u4ee3\u3002<\/p>\n<p>\u672c\u6587\uff0c\u6765\u81ea\u4e0d\u5217\u98a0\u54e5\u4f26\u6bd4\u4e9a\u5927\u5b66\u3001\u975e\u8425\u5229\u4eba\u5de5\u667a\u80fd\u7814\u7a76\u673a\u6784 Vector Institute \u7b49\u7684\u7814\u7a76\u8005\u5236\u5b9a\u4e86\u4e00\u4e2a\u65b0\u7684\u7814\u7a76\u9886\u57df\uff0c\u5373\u667a\u80fd\u4f53\u7cfb\u7edf\u7684\u81ea\u52a8\u5316\u8bbe\u8ba1\uff08ADAS\uff0cAutomated Design of Agentic Systems\uff09\uff0c\u5e76\u63d0\u51fa\u4e86\u4e00\u79cd\u7b80\u5355\u4f46\u6709\u6548\u7684 ADAS \u7b97\u6cd5\uff0c\u540d\u4e3a\u5143\u667a\u80fd\u4f53\u641c\u7d22\uff08Meta Agent Search\uff09\uff0c\u4ee5\u8bc1\u660e\u667a\u80fd\u4f53\u53ef\u4ee5\u901a\u8fc7\u4ee3\u7801\u7f16\u7a0b\u6765\u53d1\u660e\u65b0\u9896\u800c\u5f3a\u5927\u7684\u667a\u80fd\u4f53\u8bbe\u8ba1\u3002<\/p>\n<p>\u8be5\u7814\u7a76\u65e8\u5728\u81ea\u52a8\u521b\u5efa\u5f3a\u5927\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\u8bbe\u8ba1\uff0c\u5305\u62ec\u5f00\u53d1\u65b0\u7684\u6784\u5efa\u5757\u5e76\u4ee5\u65b0\u7684\u65b9\u5f0f\u7ec4\u5408\u5b83\u4eec\u3002<\/p>\n<p>\u5b9e\u9a8c\u8868\u660e\uff0c\u57fa\u4e8e ADAS \u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u7684\u6027\u80fd\u5927\u5927\u4f18\u4e8e\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u57fa\u7ebf\u3002\u4f8b\u5982\uff0c\u672c\u6587\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\u5728 DROP \u7684\u9605\u8bfb\u7406\u89e3\u4efb\u52a1\u4e2d\u5c06 F1 \u5206\u6570\u63d0\u9ad8\u4e86 13.6\/100\uff08\u4e0e\u57fa\u7ebf\u6bd4\uff09\uff0c\u5728 MGSM \u7684\u6570\u5b66\u4efb\u52a1\u4e2d\u5c06\u51c6\u786e\u7387\u63d0\u9ad8\u4e86 14.4%\u3002\u6b64\u5916\uff0c\u5728\u8de8\u57df\u8fc1\u79fb\u540e\uff0c\u5b83\u4eec\u5728 GSM8K \u548c GSM-Hard \u6570\u5b66\u4efb\u52a1\u4e0a\u7684\u51c6\u786e\u7387\u5206\u522b\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 25.9% \u548c 13.2%\u3002<\/p>\n<p>\u4e0e\u624b\u5de5\u8bbe\u8ba1\u7684\u89e3\u51b3\u65b9\u6848\u76f8\u6bd4\uff0c\u672c\u6587\u7b97\u6cd5\u8868\u73b0\u51fa\u8272\uff0c\u8fd9\u8bf4\u660e ADAS \u5728\u81ea\u52a8\u5316\u667a\u80fd\u4f53\u7cfb\u7edf\u8bbe\u8ba1\u65b9\u9762\u7684\u6f5c\u529b\u3002\u6b64\u5916\uff0c\u5b9e\u9a8c\u8868\u660e\uff0c\u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u4e0d\u4ec5\u5728\u8de8\u76f8\u4f3c\u9886\u57df\u8fc1\u79fb\u65f6\u8868\u73b0\u826f\u597d\uff0c\u800c\u4e14\u5728\u8de8\u4e0d\u540c\u9886\u57df\u8fc1\u79fb\u65f6\u4e5f\u8868\u73b0\u51fa\u8272\uff0c\u4f8b\u5982\u4ece\u6570\u5b66\u5230\u9605\u8bfb\u7406\u89e3\u3002<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/s9.51cto.com\/oss\/202408\/23\/16c18ff57c661733de25563315374eddbc6f58.webp\" data-type=\"block\" \/><\/p>\n<ul data-id=\"ua73dd4b-iS9g0g9l\">\n<li data-id=\"ld70c578-RqDqap31\">\u8bba\u6587\u5730\u5740\uff1ahttps:\/\/arxiv.org\/pdf\/2408.08435<\/li>\n<li data-id=\"ld70c578-jjmJ7APN\">\u9879\u76ee\u5730\u5740\uff1ahttps:\/\/github.com\/ShengranHu\/ADAS<\/li>\n<li data-id=\"ld70c578-FjX5AD5t\">\u8bba\u6587\u4e3b\u9875\uff1ahttps:\/\/www.shengranhu.com\/ADAS\/<\/li>\n<li data-id=\"ld70c578-I8iRJI5L\">\u8bba\u6587\u6807\u9898\uff1aAutomated Design of Agentic Systems<\/li>\n<\/ul>\n<h4>\u65b0\u7814\u7a76\u9886\u57df\uff1a\u667a\u80fd\u4f53\u7cfb\u7edf\u7684\u81ea\u52a8\u5316\u8bbe\u8ba1\uff08ADAS\uff09<\/h4>\n<p>\u8be5\u7814\u7a76\u63d0\u51fa\u4e86\u4e00\u4e2a\u65b0\u7684\u7814\u7a76\u9886\u57df \u2014\u2014 \u667a\u80fd\u4f53\u7cfb\u7edf\u7684\u81ea\u52a8\u5316\u8bbe\u8ba1\uff08Automated Design of Agentic Systems\uff0cADAS\uff09\uff0c\u5e76\u63cf\u8ff0\u4e86 ADAS \u7b97\u6cd5\u7684\u4e09\u4e2a\u5173\u952e\u7ec4\u6210\u90e8\u5206\u2014\u2014\u641c\u7d22\u7a7a\u95f4\u3001\u641c\u7d22\u7b97\u6cd5\u3001\u8bc4\u4f30\u51fd\u6570\u3002ADAS\u00a0\u4f7f\u7528\u641c\u7d22\u7b97\u6cd5\u8de8\u641c\u7d22\u7a7a\u95f4\u6765\u53d1\u73b0\u667a\u80fd\u4f53\u7cfb\u7edf\u3002<\/p>\n<p><img decoding=\"async\" title=\"\u56fe\u7247\" src=\"https:\/\/s6.51cto.com\/oss\/202408\/23\/587d74820e0523a4c71027249bd2eda28bb235.webp\" alt=\"\u56fe\u7247\" data-type=\"inline\" \/><\/p>\n<p><strong>\u641c\u7d22\u7a7a\u95f4<\/strong>\uff1a\u641c\u7d22\u7a7a\u95f4\u5b9a\u4e49\u4e86\u54ea\u4e9b\u667a\u80fd\u4f53\u7cfb\u7edf\u53ef\u4ee5\u5728 ADAS \u4e2d\u88ab\u8868\u5f81\u5e76\u88ab\u53d1\u73b0\u3002\u4f8b\u5982\uff0c\u50cf PromptBreeder (Fernando et al., 2024) \u8fd9\u6837\u7684\u5de5\u4f5c\u4ec5\u6539\u53d8\u667a\u80fd\u4f53\u7684\u6587\u672c\u63d0\u793a\uff0c\u800c\u5176\u4ed6\u7ec4\u4ef6\uff08\u4f8b\u5982\u63a7\u5236\u6d41\uff09\u4fdd\u6301\u4e0d\u53d8\u3002\u56e0\u6b64\uff0c\u5728\u641c\u7d22\u7a7a\u95f4\u4e2d\uff0c\u65e0\u6cd5\u8868\u5f81\u5177\u6709\u4e0e\u9884\u5b9a\u4e49\u63a7\u5236\u6d41\u4e0d\u540c\u63a7\u5236\u6d41\u7684\u667a\u80fd\u4f53\u3002<\/p>\n<p><strong>\u641c\u7d22\u7b97\u6cd5<\/strong>\uff1a\u641c\u7d22\u7b97\u6cd5\u5b9a\u4e49\u4e86 ADAS \u7b97\u6cd5\u5982\u4f55\u63a2\u7d22\u641c\u7d22\u7a7a\u95f4\u3002\u7531\u4e8e\u641c\u7d22\u7a7a\u95f4\u901a\u5e38\u975e\u5e38\u5927\u751a\u81f3\u65e0\u754c\uff0c\u56e0\u6b64\u5e94\u8003\u8651\u63a2\u7d22\u4e0e\u5229\u7528\u7684\u6743\u8861\uff08Sutton &amp; Barto\uff0c2018\uff09\u3002\u7406\u60f3\u60c5\u51b5\u4e0b\uff0c\u8be5\u7b97\u6cd5\u65e2\u53ef\u4ee5\u5feb\u901f\u53d1\u73b0\u9ad8\u6027\u80fd\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u53c8\u53ef\u4ee5\u907f\u514d\u9677\u5165\u5c40\u90e8\u6700\u4f18\u3002\u73b0\u6709\u65b9\u6cd5\u5305\u62ec\u4f7f\u7528\u5f3a\u5316\u5b66\u4e60\uff08Zhuge et al., 2024\uff09\u6216\u8fed\u4ee3\u751f\u6210\u65b0\u89e3\u51b3\u65b9\u6848\u7684 FM\uff08Fernando et al., 2024\uff09\u4f5c\u4e3a\u641c\u7d22\u7b97\u6cd5\u3002<\/p>\n<p><strong>\u8bc4\u4f30\u51fd\u6570<\/strong>\uff1a\u6839\u636e ADAS \u7b97\u6cd5\u7684\u5e94\u7528\uff0c\u53ef\u80fd\u9700\u8981\u8003\u8651\u4e0d\u540c\u7684\u4f18\u5316\u76ee\u6807\uff0c\u4f8b\u5982\u667a\u80fd\u4f53\u7684\u6027\u80fd\u3001\u6210\u672c\u3001\u5ef6\u8fdf\u6216\u5b89\u5168\u6027\u3002\u8bc4\u4f30\u51fd\u6570\u5b9a\u4e49\u5982\u4f55\u8bc4\u4f30\u5019\u9009\u667a\u80fd\u4f53\u7684\u8fd9\u4e9b\u6307\u6807\u3002\u4f8b\u5982\uff0c\u4e3a\u4e86\u8bc4\u4f30\u667a\u80fd\u4f53\u5728\u672a\u89c1\u8fc7\u7684\u6570\u636e\u4e0a\u7684\u8868\u73b0\uff0c\u4e00\u79cd\u7b80\u5355\u7684\u65b9\u6cd5\u662f\u8ba1\u7b97\u4efb\u52a1\u9a8c\u8bc1\u6570\u636e\u7684\u51c6\u786e\u7387\u3002<\/p>\n<p>\u8be5\u7814\u7a76\u63d0\u51fa\u7684\u7b80\u5355\u4f46\u6709\u6548\u7684 ADAS \u7b97\u6cd5\u2014\u2014\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6838\u5fc3\u6982\u5ff5\u662f\u6307\u793a\u5143\u667a\u80fd\u4f53\u8fed\u4ee3\u5730\u521b\u5efa\u6709\u8da3\u7684\u65b0\u667a\u80fd\u4f53\uff0c\u8bc4\u4f30\u5b83\u4eec\uff0c\u5c06\u5b83\u4eec\u6dfb\u52a0\u5230\u667a\u80fd\u4f53\u5b58\u50a8\u5e93\u4e2d\uff0c\u5e76\u4f7f\u7528\u6b64\u5b58\u50a8\u5e93\u5e2e\u52a9\u5143\u667a\u80fd\u4f53\u5728\u540e\u7eed\u8fed\u4ee3\u4e2d\u521b\u5efa\u66f4\u6709\u8da3\u7684\u65b0\u667a\u80fd\u4f53\u3002\u4e0e\u73b0\u6709\u7684\u5229\u7528\u4eba\u7c7b\u5174\u8da3\u6982\u5ff5\u7684\u5f00\u653e\u5f0f\u7b97\u6cd5\u7c7b\u4f3c\uff0c\u8be5\u7814\u7a76\u9f13\u52b1\u5143\u667a\u80fd\u4f53\u63a2\u7d22\u6709\u8da3\u7684\u3001\u6709\u4ef7\u503c\u7684\u667a\u80fd\u4f53\u3002<\/p>\n<p>\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6838\u5fc3\u601d\u60f3\u662f\u91c7\u7528 FM\u00a0\u4f5c\u4e3a\u641c\u7d22\u7b97\u6cd5\uff0c\u57fa\u4e8e\u4e0d\u65ad\u589e\u957f\u7684\u667a\u80fd\u4f53\u5b58\u50a8\u5e93\u6765\u8fed\u4ee3\u7f16\u7a0b\u6709\u8da3\u7684\u65b0\u667a\u80fd\u4f53\u3002\u8be5\u7814\u7a76\u4e3a\u5143\u667a\u80fd\u4f53\u5b9a\u4e49\u4e86\u4e00\u4e2a\u7b80\u5355\u7684\u6846\u67b6\uff08100 \u884c\u4ee3\u7801\u4ee5\u5185\uff09\uff0c\u4e3a\u5176\u63d0\u4f9b\u4e86\u4e00\u7ec4\u57fa\u672c\u529f\u80fd\uff0c\u4f8b\u5982\u67e5\u8be2 FM \u6216\u683c\u5f0f\u5316\u63d0\u793a\u3002<\/p>\n<p>\u56e0\u6b64\uff0c\u5143\u667a\u80fd\u4f53\u53ea\u9700\u8981\u7f16\u5199\u4e00\u4e2a\u300c\u524d\u5411\u300d\u51fd\u6570\u6765\u5b9a\u4e49\u4e00\u4e2a\u65b0\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u7c7b\u4f3c\u4e8e FunSearch \u4e2d\u7684\u505a\u6cd5\uff08Romera-Paredes et al., 2024\uff09\u3002\u8be5\u51fd\u6570\u63a5\u6536\u4efb\u52a1\u4fe1\u606f\u5e76\u8f93\u51fa\u667a\u80fd\u4f53\u5bf9\u4efb\u52a1\u7684\u54cd\u5e94\u3002<\/p>\n<p>\u5982\u56fe 1 \u6240\u793a\uff0c\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6838\u5fc3\u601d\u60f3\u662f\u8ba9\u5143\u667a\u80fd\u4f53\u5728\u4ee3\u7801\u4e2d\u8fed\u4ee3\u5730\u7f16\u7a0b\u65b0\u7684\u667a\u80fd\u4f53\u3002\u4e0b\u9762\u663e\u793a\u4e86\u5143\u667a\u80fd\u4f53\u7a0b\u5e8f\u65b0\u667a\u80fd\u4f53\u7a0b\u5e8f\u7684\u4e3b\u8981\u63d0\u793a\uff0c\u5176\u4e2d\u63d0\u793a\u4e2d\u7684\u53d8\u91cf\u9ad8\u4eae\u663e\u793a\u3002<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/s5.51cto.com\/oss\/202408\/23\/a768a8d1432113be4d7792e8b4fcc8c95d6721.webp\" data-type=\"block\" \/><\/p>\n<p><img decoding=\"async\" title=\"\u56fe\u7247\" src=\"https:\/\/s9.51cto.com\/oss\/202408\/23\/a823ed0695ce0fa3c11599e026b419ec40b16f.webp\" alt=\"\u56fe\u7247\" data-type=\"inline\" \/><\/p>\n<h4>\u5b9e\u9a8c<\/h4>\n<p>\u6240\u6709\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\u672c\u6587\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u5927\u5927\u4f18\u4e8e\u57fa\u7ebf\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u8be5\u7814\u7a76\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u5728 DROP \u9605\u8bfb\u7406\u89e3\u4efb\u52a1\u4e0a\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 13.6\/100\uff08F1 \u5206\u6570\uff09\uff0c\u5728 MGSM \u6570\u5b66\u4efb\u52a1\u4e0a\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 14.4%\uff08\u51c6\u786e\u7387\uff09\u3002\u6b64\u5916\uff0c\u7814\u7a76\u8005\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u5728\u4ece GPT-3.5 \u8fc1\u79fb\u5230 GPT-4 \u540e\uff0c\u5728 ARC \u4efb\u52a1\u4e0a\u7684\u8868\u73b0\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 14%\uff08\u51c6\u786e\u7387\uff09\uff0c\u5728\u4ece MGSM \u6570\u5b66\u4efb\u52a1\u8fc1\u79fb\u5230 GSM8K \u548c GSM-Hard \u4e2d\u7684 held-out \u6570\u5b66\u4efb\u52a1\u540e\uff0c\u51c6\u786e\u7387\u5206\u522b\u63d0\u9ad8\u4e86 25.9% \u548c 13.2%\u3002<\/p>\n<p><strong>\u6848\u4f8b\u7814\u7a76\uff1aARC \u6311\u6218<\/strong><\/p>\n<p>\u5982\u56fe 3a \u6240\u793a\uff0c\u5143\u667a\u80fd\u4f53\u641c\u7d22\u53ef\u4ee5\u6709\u6548\u4e14\u9010\u6b65\u5730\u53d1\u73b0\u6027\u80fd\u4f18\u4e8e\u6700\u65b0\u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\u3002\u6587\u672c\u6846\u4e2d\u7a81\u51fa\u663e\u793a\u4e86\u91cd\u8981\u7684\u7a81\u7834\u3002<\/p>\n<p>\u6b64\u5916\uff0c\u56fe 3b \u663e\u793a\u4e86\u53d1\u73b0\u6700\u597d\u7684\u667a\u80fd\u4f53\uff0c\u5176\u4e2d\u91c7\u7528\u4e86\u590d\u6742\u7684\u53cd\u9988\u673a\u5236\u6765\u66f4\u6709\u6548\u5730\u7ec6\u5316\u7b54\u6848\u3002\u4ed4\u7ec6\u89c2\u5bdf\u641c\u7d22\u8fdb\u5ea6\u5c31\u4f1a\u53d1\u73b0\uff0c\u8fd9\u79cd\u590d\u6742\u7684\u53cd\u9988\u673a\u5236\u5e76\u4e0d\u662f\u7a81\u7136\u51fa\u73b0\u7684\u3002<\/p>\n<p><img decoding=\"async\" title=\"\u56fe\u7247\" src=\"https:\/\/s7.51cto.com\/oss\/202408\/23\/08bde12501a7431219c880fb3b3564e9ff5d98.webp\" alt=\"\u56fe\u7247\" data-type=\"inline\" \/><\/p>\n<p><strong>\u63a8\u7406\u548c\u95ee\u9898 &#8211; \u89e3\u51b3\u57df<\/strong><\/p>\n<p>\u8de8\u591a\u4e2a\u57df\u7684\u7ed3\u679c\u8868\u660e\uff0c\u5143\u667a\u80fd\u4f53\u641c\u7d22\u53ef\u4ee5\u53d1\u73b0\u8868\u73b0\u4f18\u4e8e SOTA \u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\uff08\u8868 1\uff09\u3002<\/p>\n<p><img decoding=\"async\" title=\"\u56fe\u7247\" src=\"https:\/\/s2.51cto.com\/oss\/202408\/23\/067faa14137157acf01113f29fa544426bbb6f.webp\" alt=\"\u56fe\u7247\" data-type=\"inline\" \/><\/p>\n<p><strong>\u6cdb\u5316\u6027\u4ee5\u53ca\u53ef\u8fc1\u79fb\u6027<\/strong><\/p>\n<p>\u7814\u7a76\u8005\u8fdb\u4e00\u6b65\u5c55\u793a\u4e86\u6240\u53d1\u73b0\u667a\u80fd\u4f53\u7684\u53ef\u8fc1\u79fb\u6027\u548c\u53ef\u6cdb\u5316\u6027\u3002<\/p>\n<p>\u5982\u8868 2 \u6240\u793a\uff0c\u7814\u7a76\u8005\u89c2\u5bdf\u5230\u641c\u7d22\u5230\u7684\u667a\u80fd\u4f53\uff08searched agent\uff09\u59cb\u7ec8\u4f18\u4e8e\u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\uff0c\u5e76\u4e14\u5dee\u8ddd\u5f88\u5927\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u7814\u7a76\u8005\u53d1\u73b0 Anthropic \u6700\u5f3a\u5927\u7684\u6a21\u578b Claude-Sonnet \u5728\u6240\u6709\u6d4b\u8bd5\u6a21\u578b\u4e2d\u8868\u73b0\u6700\u4f73\uff0c\u4f7f\u57fa\u4e8e\u8be5\u6a21\u578b\u7684\u667a\u80fd\u4f53\u5728 ARC \u4e0a\u5b9e\u73b0\u4e86\u8fd1 50% \u7684\u51c6\u786e\u7387\u3002<\/p>\n<p><img decoding=\"async\" title=\"\u56fe\u7247\" src=\"https:\/\/s3.51cto.com\/oss\/202408\/23\/e437b67044c668057d47760ca73a62395aa758.webp\" alt=\"\" data-type=\"inline\" \/><\/p>\n<p>\u5982\u8868 3 \u6240\u793a\uff0c\u7814\u7a76\u8005\u89c2\u5bdf\u5230\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6027\u80fd\u4e0e\u57fa\u7ebf\u76f8\u6bd4\u5177\u6709\u7c7b\u4f3c\u7684\u4f18\u52bf\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u4e0e\u57fa\u7ebf\u76f8\u6bd4\uff0c\u672c\u6587\u7684\u667a\u80fd\u4f53\u5728 GSM8K \u548c GSM-Hard \u4e0a\u7684\u51c6\u786e\u7387\u5206\u522b\u63d0\u9ad8\u4e86 25.9% \u548c 13.2%\u3002<\/p>\n<p><img decoding=\"async\" title=\"\u56fe\u7247\" src=\"https:\/\/s6.51cto.com\/oss\/202408\/23\/4814d7a680ad36f0c4a034b9fc1805ebe81052.webp\" alt=\"\u56fe\u7247\" data-type=\"inline\" \/><\/p>\n<p>\u66f4\u4ee4\u4eba\u60ca\u8bb6\u7684\u662f\uff0c\u7814\u7a76\u8005\u89c2\u5bdf\u5230\u5728\u6570\u5b66\u9886\u57df\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u53ef\u4ee5\u8fc1\u79fb\u5230\u5230\u975e\u6570\u5b66\u9886\u57df\uff08\u8868 4\uff09\u3002<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/s8.51cto.com\/oss\/202408\/23\/c6c418581fa7ce1d1065605d04ee9ca696ed8e.webp\" data-type=\"block\" \/><\/p>\n<\/div>\n<\/div>\n<div class=\"pvc_clear\"><\/div>\n<p id=\"pvc_stats_21855\" class=\"pvc_stats total_only  \" data-element-id=\"21855\" style=\"\"><i class=\"pvc-stats-icon medium\" aria-hidden=\"true\"><svg xmlns=\"http:\/\/www.w3.org\/2000\/svg\" version=\"1.0\" viewBox=\"0 0 502 315\" preserveAspectRatio=\"xMidYMid meet\"><g transform=\"translate(0,332) scale(0.1,-0.1)\" fill=\"\" stroke=\"none\"><path d=\"M2394 3279 l-29 -30 -3 -207 c-2 -182 0 -211 15 -242 39 -76 157 -76 196 0 15 31 17 60 15 243 l-3 209 -33 29 c-26 23 -41 29 -80 29 -41 0 -53 -5 -78 -31z\"\/><path d=\"M3085 3251 c-45 -19 -58 -50 -96 -229 -47 -217 -49 -260 -13 -295 52 -53 146 -42 177 20 16 31 87 366 87 410 0 70 -86 122 -155 94z\"\/><path d=\"M1751 3234 c-13 -9 -29 -31 -37 -50 -12 -29 -10 -49 21 -204 19 -94 39 -189 45 -210 14 -50 54 -80 110 -80 34 0 48 6 76 34 21 21 34 44 34 59 0 14 -18 113 -40 219 -37 178 -43 195 -70 221 -36 32 -101 37 -139 11z\"\/><path d=\"M1163 3073 c-36 -7 -73 -59 -73 -102 0 -56 133 -378 171 -413 34 -32 83 -37 129 -13 70 36 67 87 -16 290 -86 209 -89 214 -129 231 -35 14 -42 15 -82 7z\"\/><path d=\"M3689 3066 c-15 -9 -33 -30 -42 -48 -48 -103 -147 -355 -147 -375 0 -98 131 -148 192 -74 13 15 57 108 97 206 80 196 84 226 37 273 -30 30 -99 39 -137 18z\"\/><path d=\"M583 2784 c-38 -19 -67 -74 -58 -113 9 -42 211 -354 242 -373 16 -10 45 -18 66 -18 51 0 107 52 107 100 0 39 -1 41 -124 234 -80 126 -108 162 -133 173 -41 17 -61 16 -100 -3z\"\/><path d=\"M4250 2784 c-14 -9 -74 -91 -133 -183 -95 -150 -107 -173 -107 -213 0 -55 33 -94 87 -104 67 -13 90 8 211 198 130 202 137 225 78 284 -27 27 -42 34 -72 34 -22 0 -50 -8 -64 -16z\"\/><path d=\"M2275 2693 c-553 -48 -1095 -270 -1585 -649 -135 -104 -459 -423 -483 -476 -23 -49 -22 -139 2 -186 73 -142 361 -457 571 -626 285 -228 642 -407 990 -497 242 -63 336 -73 660 -74 310 0 370 5 595 52 535 111 1045 392 1455 803 122 121 250 273 275 326 19 41 19 137 0 174 -41 79 -309 363 -465 492 -447 370 -946 591 -1479 653 -113 14 -422 18 -536 8z m395 -428 c171 -34 330 -124 456 -258 112 -119 167 -219 211 -378 27 -96 24 -300 -5 -401 -72 -255 -236 -447 -474 -557 -132 -62 -201 -76 -368 -76 -167 0 -236 14 -368 76 -213 98 -373 271 -451 485 -162 444 86 934 547 1084 153 49 292 57 452 25z m909 -232 c222 -123 408 -262 593 -441 76 -74 138 -139 138 -144 0 -16 -233 -242 -330 -319 -155 -123 -309 -223 -461 -299 l-81 -41 32 46 c18 26 49 83 70 128 143 306 141 649 -6 957 -25 52 -61 116 -79 142 l-34 47 45 -20 c26 -10 76 -36 113 -56z m-2057 25 c-40 -58 -105 -190 -130 -263 -110 -324 -59 -707 132 -981 25 -35 42 -64 37 -64 -19 0 -241 119 -326 174 -188 122 -406 314 -532 468 l-58 71 108 103 c185 178 428 349 672 473 66 33 121 60 123 61 2 0 -10 -19 -26 -42z\"\/><path d=\"M2375 1950 c-198 -44 -350 -190 -395 -379 -18 -76 -8 -221 19 -290 114 -284 457 -406 731 -260 98 52 188 154 231 260 27 69 37 214 19 290 -38 163 -166 304 -326 360 -67 23 -215 33 -279 19z\"\/><\/g><\/svg><\/i> <img loading=\"lazy\" decoding=\"async\" width=\"16\" height=\"16\" alt=\"Loading\" src=\"https:\/\/aif.amtbbs.org\/wp-content\/plugins\/page-views-count\/ajax-loader-2x.gif\" border=0 \/><\/p>\n<div class=\"pvc_clear\"><\/div>\n","protected":false},"excerpt":{"rendered":"<p>\u57fa\u4e8e ADAS \u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u7684\u6027\u80fd\u5927\u5927\u4f18\u4e8e\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u57fa\u7ebf\u3002 \u57fa\u7840\u6a21\u578b (FM) \u5982 GPT \u548c Cl [&hellip;]<\/p>\n<div class=\"pvc_clear\"><\/div>\n<p id=\"pvc_stats_21855\" class=\"pvc_stats total_only  \" data-element-id=\"21855\" style=\"\"><i class=\"pvc-stats-icon medium\" aria-hidden=\"true\"><svg xmlns=\"http:\/\/www.w3.org\/2000\/svg\" version=\"1.0\" viewBox=\"0 0 502 315\" preserveAspectRatio=\"xMidYMid meet\"><g transform=\"translate(0,332) scale(0.1,-0.1)\" fill=\"\" stroke=\"none\"><path d=\"M2394 3279 l-29 -30 -3 -207 c-2 -182 0 -211 15 -242 39 -76 157 -76 196 0 15 31 17 60 15 243 l-3 209 -33 29 c-26 23 -41 29 -80 29 -41 0 -53 -5 -78 -31z\"\/><path d=\"M3085 3251 c-45 -19 -58 -50 -96 -229 -47 -217 -49 -260 -13 -295 52 -53 146 -42 177 20 16 31 87 366 87 410 0 70 -86 122 -155 94z\"\/><path d=\"M1751 3234 c-13 -9 -29 -31 -37 -50 -12 -29 -10 -49 21 -204 19 -94 39 -189 45 -210 14 -50 54 -80 110 -80 34 0 48 6 76 34 21 21 34 44 34 59 0 14 -18 113 -40 219 -37 178 -43 195 -70 221 -36 32 -101 37 -139 11z\"\/><path d=\"M1163 3073 c-36 -7 -73 -59 -73 -102 0 -56 133 -378 171 -413 34 -32 83 -37 129 -13 70 36 67 87 -16 290 -86 209 -89 214 -129 231 -35 14 -42 15 -82 7z\"\/><path d=\"M3689 3066 c-15 -9 -33 -30 -42 -48 -48 -103 -147 -355 -147 -375 0 -98 131 -148 192 -74 13 15 57 108 97 206 80 196 84 226 37 273 -30 30 -99 39 -137 18z\"\/><path d=\"M583 2784 c-38 -19 -67 -74 -58 -113 9 -42 211 -354 242 -373 16 -10 45 -18 66 -18 51 0 107 52 107 100 0 39 -1 41 -124 234 -80 126 -108 162 -133 173 -41 17 -61 16 -100 -3z\"\/><path d=\"M4250 2784 c-14 -9 -74 -91 -133 -183 -95 -150 -107 -173 -107 -213 0 -55 33 -94 87 -104 67 -13 90 8 211 198 130 202 137 225 78 284 -27 27 -42 34 -72 34 -22 0 -50 -8 -64 -16z\"\/><path d=\"M2275 2693 c-553 -48 -1095 -270 -1585 -649 -135 -104 -459 -423 -483 -476 -23 -49 -22 -139 2 -186 73 -142 361 -457 571 -626 285 -228 642 -407 990 -497 242 -63 336 -73 660 -74 310 0 370 5 595 52 535 111 1045 392 1455 803 122 121 250 273 275 326 19 41 19 137 0 174 -41 79 -309 363 -465 492 -447 370 -946 591 -1479 653 -113 14 -422 18 -536 8z m395 -428 c171 -34 330 -124 456 -258 112 -119 167 -219 211 -378 27 -96 24 -300 -5 -401 -72 -255 -236 -447 -474 -557 -132 -62 -201 -76 -368 -76 -167 0 -236 14 -368 76 -213 98 -373 271 -451 485 -162 444 86 934 547 1084 153 49 292 57 452 25z m909 -232 c222 -123 408 -262 593 -441 76 -74 138 -139 138 -144 0 -16 -233 -242 -330 -319 -155 -123 -309 -223 -461 -299 l-81 -41 32 46 c18 26 49 83 70 128 143 306 141 649 -6 957 -25 52 -61 116 -79 142 l-34 47 45 -20 c26 -10 76 -36 113 -56z m-2057 25 c-40 -58 -105 -190 -130 -263 -110 -324 -59 -707 132 -981 25 -35 42 -64 37 -64 -19 0 -241 119 -326 174 -188 122 -406 314 -532 468 l-58 71 108 103 c185 178 428 349 672 473 66 33 121 60 123 61 2 0 -10 -19 -26 -42z\"\/><path d=\"M2375 1950 c-198 -44 -350 -190 -395 -379 -18 -76 -8 -221 19 -290 114 -284 457 -406 731 -260 98 52 188 154 231 260 27 69 37 214 19 290 -38 163 -166 304 -326 360 -67 23 -215 33 -279 19z\"\/><\/g><\/svg><\/i> <img loading=\"lazy\" decoding=\"async\" width=\"16\" height=\"16\" alt=\"Loading\" src=\"https:\/\/aif.amtbbs.org\/wp-content\/plugins\/page-views-count\/ajax-loader-2x.gif\" border=0 \/><\/p>\n<div class=\"pvc_clear\"><\/div>\n","protected":false},"author":56,"featured_media":21857,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[23,80],"tags":[124,1275],"class_list":["post-21855","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-23","category-80","tag-ai","tag-1275"],"_links":{"self":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/21855","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/users\/56"}],"replies":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/comments?post=21855"}],"version-history":[{"count":1,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/21855\/revisions"}],"predecessor-version":[{"id":21858,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/21855\/revisions\/21858"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media\/21857"}],"wp:attachment":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media?parent=21855"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/categories?post=21855"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/tags?post=21855"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}