{"id":5126,"date":"2026-05-22T21:00:08","date_gmt":"2026-05-22T13:00:08","guid":{"rendered":"https:\/\/1ow.top\/index.php\/2026\/05\/22\/%e6%9d%8e%e9%a3%9e%e9%a3%9e%e5%86%8d%e5%87%ba%e6%89%8b%ef%bc%8c%e7%a9%ba%e9%97%b4%e6%99%ba%e8%83%bd%e7%9a%84imagenet%e6%9d%a5%e4%ba%86\/"},"modified":"2026-05-22T21:00:08","modified_gmt":"2026-05-22T13:00:08","slug":"%e6%9d%8e%e9%a3%9e%e9%a3%9e%e5%86%8d%e5%87%ba%e6%89%8b%ef%bc%8c%e7%a9%ba%e9%97%b4%e6%99%ba%e8%83%bd%e7%9a%84imagenet%e6%9d%a5%e4%ba%86","status":"publish","type":"post","link":"https:\/\/1ow.top\/index.php\/2026\/05\/22\/%e6%9d%8e%e9%a3%9e%e9%a3%9e%e5%86%8d%e5%87%ba%e6%89%8b%ef%bc%8c%e7%a9%ba%e9%97%b4%e6%99%ba%e8%83%bd%e7%9a%84imagenet%e6%9d%a5%e4%ba%86\/","title":{"rendered":"\u674e\u98de\u98de\u518d\u51fa\u624b\uff0c\u7a7a\u95f4\u667a\u80fd\u7684ImageNet\u6765\u4e86"},"content":{"rendered":"
\"\u674e\u98de\u98de\u518d\u51fa\u624b\uff0c\u7a7a\u95f4\u667a\u80fd\u7684ImageNet\u6765\u4e86\"\/<\/div><\/figure>\n
\"\u674e\u98de\u98de\u518d\u51fa\u624b\uff0c\u7a7a\u95f4\u667a\u80fd\u7684ImageNet\u6765\u4e86\"\/<\/div><\/figure>\n
\"\u674e\u98de\u98de\u518d\u51fa\u624b\uff0c\u7a7a\u95f4\u667a\u80fd\u7684ImageNet\u6765\u4e86\"\/<\/div><\/figure>\n
\"\u674e\u98de\u98de\u518d\u51fa\u624b\uff0c\u7a7a\u95f4\u667a\u80fd\u7684ImageNet\u6765\u4e86\"\/<\/div><\/figure>\n

\u4e00\u4e2a\u4e13\u95e8\u7528\u6765\u8bc4\u6d4b\u5177\u8eab\u7a7a\u95f4\u667a\u80fd\u7684\u65b0\u57fa\u51c6<\/p>\n

\u542c\u96e8 \u53d1\u81ea \u51f9\u975e\u5bfa<\/p>\n

ImageNet\u4e4b\u540e\uff0c\u674e\u98de\u98de\u518d\u51fa\u624b\uff01<\/p>\n

\u674e\u98de\u98de\u56e2\u961f\u6700\u65b0\u53d1\u5e03ESI-Bench\u2014\u2014\u4e00\u4e2a\u4e13\u95e8\u7528\u6765\u8bc4\u6d4b\u5177\u8eab\u7a7a\u95f4\u667a\u80fd\u7684\u65b0\u57fa\u51c6\u3002<\/p>\n

\u8fc7\u53bb\u7684\u7a7a\u95f4\u667a\u80fd\u8bc4\u6d4b\u9ed8\u8ba4\u7ed9\u6a21\u578b\u6700\u4f18\u89c2\u6d4b\uff0c\u800cESI-Bench\u7b2c\u4e00\u4e2a\u628a\u89c2\u5bdf\u8005\u53d8\u6210\u884c\u52a8\u8005\uff0c\u95ed\u5408\u4e86\u611f\u77e5-\u884c\u52a8\u56de\u8def\u3002<\/p>\n

\u5b83\u4e3a\u5177\u8eab\u7a7a\u95f4\u667a\u80fd\u9886\u57df\u63d0\u4f9b\u4e86\u4e00\u4e2a\u7cfb\u7edf\u6027\u7684\u8bc4\u6d4b\u6846\u67b6\uff0c\u8986\u76d6\u4eba\u7c7b\u6838\u5fc3\u7a7a\u95f4\u8ba4\u77e5\u80fd\u529b\u7684\u56db\u5927\u7ef4\u5ea6\u3002<\/p>\n

\u8bba\u6587\u7684\u6838\u5fc3\u7ed3\u8bba\u662f\uff1a\u73b0\u5728\u7684AI\u770b\u56fe\u5f88\u5389\u5bb3\uff0c\u4f46\u79bb\u300c\u4f1a\u52a8\u3001\u4f1a\u6478\u3001\u4f1a\u4e3b\u52a8\u627e\u7b54\u6848\u300d\u7684\u7a7a\u95f4\u667a\u80fd\u8fd8\u5dee\u5f97\u8fdc\u3002<\/p>\n

ESI-Bench\u53d1\u5e03\u7684\u80cc\u666f\uff0c\u662f\u7531\u4e8e\u76ee\u524d\u7684\u7a7a\u95f4\u667a\u80fdbenchmark\uff0c\u6d4b\u7684\u90fd\u662f\u300c\u88ab\u52a8\u611f\u77e5\u300d\u3002<\/p>\n

\u628a\u4e00\u5f20\u6216\u51e0\u5f20\u56fe\u7247\u6254\u7ed9\u6a21\u578b\uff0c\u95ee\u300cA\u7269\u4f53\u5728B\u7269\u4f53\u7684\u5de6\u8fb9\u8fd8\u662f\u53f3\u8fb9\u300d\u300c\u8fd9\u4e2a\u676f\u5b50\u80fd\u88c5\u591a\u5c11\u6c34\u300d\u300c\u62bd\u5c49\u91cc\u6709\u6ca1\u6709\u4e1c\u897f\u300d\uff0c\u8fd9\u6837\u7684\u9898\u76ee\u6d4b\u51fa\u6765\u7684\u662f\u6a21\u578b\u7684\u89c6\u529b\uff0c\u800c\u975e\u7a7a\u95f4\u63a8\u7406\u80fd\u529b\u3002<\/p>\n

\u53cd\u89c2\u4eba\u7c7b\u662f\u600e\u4e48\u505a\u7684\uff1f\u4eba\u7c7b\u4f1a\u7ad9\u8d77\u6765\u7ed5\u5230\u7269\u4f53\u80cc\u540e\u53bb\u770b\uff0c\u4f1a\u628a\u62bd\u5c49\u62c9\u5f00\uff0c\u4f1a\u628a\u6c34\u5012\u51fa\u6765\u91cf\u4e00\u91cf\u3002<\/p>\n

\u8fd9\u5c31\u662fESI-Bench\u7684\u6838\u5fc3\u7acb\u573a\uff1a\u628a\u89c2\u5bdf\u8005\u53d8\u6210\u884c\u52a8\u8005\u3002<\/p>\n

\u73b0\u5b9e\u4e16\u754c\u91cc\uff0c\u667a\u80fd\u4f53\u5fc5\u987b\u50cf\u4eba\u7c7b\u4e00\u6837\uff0c\u4e3b\u52a8\u51b3\u5b9a\u884c\u52a8\u3001\u83b7\u53d6\u8bc1\u636e\uff0c\u518d\u57fa\u4e8e\u65b0\u89c2\u6d4b\u505a\u4e0b\u4e00\u6b65\u5224\u65ad\u3002\u56e2\u961f\u628a\u5b83\u79f0\u4e3a\u300c\u611f\u77e5-\u884c\u52a8\u56de\u8def\u300d\uff08Perception-Action Loop\uff09\u3002<\/p>\n

ESI-Bench\u5c31\u662f\u8fd9\u6837\u4e00\u5957\u8d85\u8d8a\u73b0\u6709\u57fa\u51c6\u7684\u7a7a\u95f4\u667a\u80fd\u65b0\u8bc4\u6d4b\u57fa\u51c6\uff0c\u5b83\u5305\u542b10\u4e2a\u4efb\u52a1\u7c7b\u522b\uff0c29\u4e2a\u5b50\u7c7b\u522b\uff0c3081\u4e2a\u4efb\u52a1\u5b9e\u4f8b\uff0c\u5168\u90e8\u5728OmniGibson\u4eff\u771f\u5e73\u53f0\u4e0a\u6784\u5efa\uff0c\u573a\u666f\u7d20\u6750\u6765\u81eaBEHAVIOR-1K\u573a\u666f\u5e93\u3002<\/p>\n

\u6240\u6709\u4efb\u52a1\u56f4\u7ed5Spelke\u7684\u56db\u5927\u6838\u5fc3\u77e5\u8bc6\u7cfb\u7edf\u8bbe\u8ba1\uff0c\u4e5f\u5c31\u662f\u4eba\u7c7b\u5a74\u513f\u5929\u751f\u5c31\u5177\u5907\u7684\u7a7a\u95f4\u76f4\u89c9\uff1a\u7269\u4f53\u8868\u5f81\u3001\u5e03\u5c40\u4e0e\u51e0\u4f55\u3001\u6570\u91cf\u8868\u5f81\u3001\u76ee\u6807\u5bfc\u5411\u884c\u52a8\u3002<\/p>\n

\u5b83\u7684\u5173\u952e\u8bbe\u5b9a\u5728\u4e8e\u884c\u52a8\u5f3a\u5236\u3002\u6bcf\u4e00\u9053\u9898\uff0cAI\u667a\u80fd\u4f53\u5fc5\u987b\u4e3b\u52a8\u884c\u52a8\u624d\u80fd\u62ff\u5230\u8db3\u591f\u4fe1\u606f\u4f5c\u7b54\u3002\u6a21\u578b\u4e0d\u80fd\u5750\u5728\u539f\u5730\u7b49\u56fe\u7247\uff0c\u5b83\u8981\u51b3\u5b9a\u5f80\u54ea\u8d70\u3001\u770b\u4ec0\u4e48\u3001\u62ff\u4ec0\u4e48\u3001\u600e\u4e48\u64cd\u4f5c\u3002<\/p>\n

\u4e3e\u51e0\u4e2a\u5177\u4f53\u7684\u4f8b\u5b50\uff1a<\/p>\n

\u6bd4\u5982\u8bc4\u6d4b\u4e2d\u6709\u4e00\u9053\u300c\u521a\u6027\u5bb9\u7eb3\u300d\u9898\uff1a\u7ed9\u5b9a\u51e0\u4e2a\u5bb9\u5668\u548c\u51e0\u4e2a\u7269\u4f53\uff0c\u8981\u6c42\u628a\u7269\u4f53\u5168\u90e8\u88c5\u8fdb\u53bb\u3002\u6709\u7684\u5bb9\u5668\u5f00\u53e3\u5c0f\u3001\u6709\u7684\u5185\u90e8\u6709\u9694\u677f\u3001\u6709\u7684\u76d6\u5b50\u9700\u8981\u6380\u5f00\u624d\u80fd\u770b\u5230\u771f\u5b9e\u5bb9\u91cf\u3002<\/p>\n

\u6a21\u578b\u5fc5\u987b\u8d70\u8fd1\u3001\u4fef\u8eab\u3001\u751a\u81f3\u628a\u5bb9\u5668\u62ff\u8d77\u6765\u4ece\u5e95\u90e8\u89c2\u5bdf\uff0c\u624d\u80fd\u5224\u65ad\u80fd\u4e0d\u80fd\u88c5\u5f97\u4e0b\u3002<\/p>\n

\u8fd8\u6709\u300c\u6db2\u4f53\u4f53\u79ef\u300d\u9898\uff1a\u4e24\u4e2a\u676f\u5b50\uff0c\u4ece\u5916\u89c2\u770b\u4e0d\u51fa\u5bb9\u91cf\u5dee\u5f02\uff0c\u6a21\u578b\u9700\u8981\u628a\u6c34\u5012\u8fdb\u53bb\u6d4b\u8bd5\uff0c\u6216\u8005\u76f4\u63a5\u62ff\u8d77\u6765\u6382\u91cf\u3002<\/p>\n

\u8fd9\u4e48\u4e00\u8bf4\uff0c\u5927\u5bb6\u5e94\u8be5\u4e5f\u80fd\u76f4\u89c2\u611f\u53d7\u5230\u8fd9\u5957\u8bc4\u6d4b\u57fa\u51c6\u7684\u8bbe\u8ba1\u7406\u5ff5\uff1a<\/p>\n

\u6b63\u786e\u7b54\u6848\u4e0d\u5728\u4efb\u4f55\u5355\u5f20\u56fe\u7247\u91cc\uff0c\u667a\u80fd\u4f53\u5fc5\u987b\u4e3b\u52a8\u884c\u52a8\u5e76\u63a8\u7406\u51fa\u6b63\u786e\u7ed3\u679c\u3002<\/p>\n

\u56e2\u961f\u7279\u522b\u6307\u51fa\uff0c\u4e0e\u6b64\u524d\u5de5\u4f5c\u76f8\u6bd4\uff0cESI-Bench\u5728\u4e09\u4e2a\u5730\u65b9\u6709\u6240\u8d85\u8d8a\uff1a<\/p>\n

\u4ece\u7a7a\u95f4\u611f\u77e5\u5230\u7a7a\u95f4\u80fd\u529b\uff1a\u5728\u8fd9\u91cc\uff0c\u667a\u80fd\u4f53\u4e0d\u4ec5\u6839\u636e\u4ed6\u4eec\u80fd\u611f\u77e5\u5230\u4ec0\u4e48\u6765\u8bc4\u4f30\uff0c\u8fd8\u6839\u636e\u4ed6\u4eec\u662f\u5426\u77e5\u9053\u90e8\u7f72\u54ea\u4e9b\u5177\u4f53\u80fd\u529b\u6765\u89e3\u51b3\u7a7a\u95f4\u4efb\u52a1\u6765\u8bc4\u4f30\uff1b<\/p>\n

\u9009\u62e9\u6027\u611f\u77e5\uff1a\u667a\u80fd\u4f53\u5fc5\u987b\u786e\u5b9a\u54ea\u4e9b\u89c2\u5bdf\u503c\u5f97\u83b7\u53d6\uff0c\u4f18\u5148\u8003\u8651\u4e0e\u4efb\u52a1\u76f8\u5173\u7684\u4fe1\u606f\u800c\u4e0d\u662f\u5197\u4f59\u6216\u65e0\u4fe1\u606f\u7684\u8f93\u5165\uff1b<\/p>\n

\u89e3\u51b3\u611f\u77e5\u6b67\u4e49\uff1a\u667a\u80fd\u4f53\u5fc5\u987b\u901a\u8fc7\u8bef\u5bfc\u6027\u89c2\u5bdf\u8fdb\u884c\u63a8\u7406\uff0c\u4ee5\u63a8\u65ad\u9690\u85cf\u7684\u7a7a\u95f4\u7ed3\u6784\u548c\u8d85\u8d8a\u76f4\u63a5\u89c2\u5bdf\u7684\u6f5c\u5728\u7269\u7406\u7ea6\u675f\u3002<\/p>\n

\u56e2\u961f\u62ff\u5f53\u524d\u6700\u5f3a\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b\u505a\u4e86\u5168\u9762\u6d4b\u8bd5\uff0c\u5305\u62ecGPT-5\u548cGemini\u7cfb\u5217\u3002<\/p>\n

\u8fd9\u662f\u6700\u4e3b\u8981\u7684\u5b9e\u9a8c\u7ed3\u679c\u56fe\uff0c\u5305\u542b\u4e86ESI-Bench\u5728\u88ab\u52a8\u611f\u77e5\u3001\u4e3b\u52a8\u63a2\u7d22\u3001Oracle\u4e09\u79cd\u8303\u5f0f\u4e0b\u7684\u5404\u9879\u4efb\u52a1\u51c6\u786e\u7387\uff0c\u6db5\u76d62D+VLM\u30013D+LLM\u53ca\u4eba\u7c7b\u57fa\u7ebf\u3002<\/p>\n

\u6838\u5fc3\u7ed3\u8bba\u67093\u4e2a\u3002<\/p>\n

\u7b2c\u4e00\uff0c\u611f\u77e5\u4e0d\u662f\u74f6\u9888\uff0c\u884c\u52a8\u624d\u662f\u3002<\/p>\n

\u597d\u6d88\u606f\u662f\uff0c\u4e3b\u52a8\u63a2\u7d22\u786e\u5b9e\u6709\u6548\u3002\u5728\u6ca1\u6709\u989d\u5916\u6307\u4ee4\u7684\u60c5\u51b5\u4e0b\uff0c\u667a\u80fd\u4f53\u81ea\u53d1\u6d8c\u73b0\u51fa\u591a\u79cd\u7a7a\u95f4\u7b56\u7565\u3002<\/p>\n

\u6bd4\u5982\u7ed5\u5230\u7269\u4f53\u80cc\u540e\u89c2\u5bdf\uff08move-behind\uff09\u3001\u5207\u6362\u4fef\u89c6\u89d2\u5ea6\uff08top-down\uff09\u3001\u628a\u7269\u4f53\u62ff\u8d77\u6765\uff08pick-up\uff09\u3001\u628a\u6c34\u5012\u51fa\u6765\u9a8c\u8bc1\uff08pour-out\uff09\u3002<\/p>\n

Gemini 3.1\u5728\u300c\u90e8\u5206\u906e\u6321\u300d\u4efb\u52a1\u4e0a\uff0c\u5982\u679c\u7ed9\u5230\u6700\u4f73\u89c2\u5bdf\u89c6\u89d2\uff0c\u51c6\u786e\u7387\u4ece14.6%\u66b4\u6da8\u523095.1%\u3002<\/p>\n

\u8fd9\u8bf4\u660e\uff0c\u6a21\u578b\u672c\u8eab\u7684\u611f\u77e5\u80fd\u529b\u662f\u597d\u7684\uff0c\u53ea\u8981\u7ed9\u5bf9\u89c6\u89d2\uff0c\u5b83\u5c31\u80fd\u770b\u5f97\u61c2\u3002<\/p>\n

\u4f46\u95ee\u9898\u5728\u4e8e\uff0c\u6a21\u578b\u81ea\u5df1\u627e\u4e0d\u5230\u90a3\u4e2a\u5bf9\u7684\u89c6\u89d2\u3002<\/p>\n

\u66f4\u7cdf\u7cd5\u7684\u95ee\u9898\u5728\u4e8e\uff0c\u88ab\u52a8\u591a\u89c6\u89d2\u7b56\u7565\u4e0d\u4ec5\u6ca1\u7528\uff0c\u53cd\u800c\u6709\u5bb3\u3002<\/p>\n

\u8ba9GPT-5\u591a\u770b\u51e0\u5f20\u968f\u673a\u89d2\u5ea6\u7684\u56fe\u7247\uff0c\u7a7a\u95f4\u8ddd\u79bb\u4efb\u52a1\u7684\u51c6\u786e\u7387\u4ece53.9%\u964d\u523049.1%\u3002\u56fe\u770b\u5f97\u591a\u4e86\uff0c\u5206\u53cd\u800c\u4f4e\u4e86\u3002<\/p>\n

\u25b3GPT-5\u548cGemini 3.1\u5728\u4e3b\u52a8\u63a2\u7d22\u4e2d\u8fbe\u5230\u6b63\u786e\u7b54\u6848\u6240\u9700\u7684\u5e73\u5747\u6b65\u6570<\/p>\n

\u56e2\u961f\u628a\u8fd9\u4e2a\u73b0\u8c61\u547d\u540d\u4e3a\u300c\u52a8\u4f5c\u76f2\u89c6\u300d\uff08Action Blindness\uff09\uff0c\u4e00\u4e2a\u5dee\u52a8\u4f5c\u5bfc\u81f4\u4e00\u4e2a\u5dee\u89c6\u89d2\uff0c\u5dee\u89c6\u89d2\u89e6\u53d1\u66f4\u5dee\u52a8\u4f5c\uff0c\u5f62\u6210\u4e0d\u53ef\u9006\u7684\u7ea7\u8054\u5931\u8d25\u3002<\/p>\n

\u5728\u7ed3\u6784\u56f4\u5408\u4efb\u52a1\u4e0a\uff0c\u4e3b\u52a8\u63a2\u7d22\u7b56\u7565\u548c\u4e0a\u5e1d\u89c6\u89d2\u7684\u5dee\u8ddd\u9ad8\u8fbe49.7%\u3002<\/p>\n

\u4e5f\u5c31\u662f\u8bf4\uff0c\u7a7a\u95f4\u667a\u80fd\u7684\u5361\u70b9\u4e0d\u5728\u4e8e\u89c6\u89c9\u6a21\u578b\u4e0d\u591f\u5f3a\uff0c\u800c\u662f\u884c\u52a8\u7b56\u7565\u51e0\u4e4e\u4e3a\u96f6\u3002<\/p>\n

\u7b2c\u4e8c\uff0c3D\u91cd\u5efa\u4e0d\u662f\u4e07\u80fd\u836f\uff0c\u4e0d\u5b8c\u7f8e\u76843D\u6bd42D\u66f4\u5751\u3002<\/p>\n

\u65e2\u71362D\u88ab\u52a8\u770b\u56fe\u4e0d\u884c\uff0c\u90a3\u4e0a3D\u5462\uff1f\u8fd9\u4e5f\u662f\u5f53\u524d\u5f88\u591a\u5177\u8eab\u667a\u80fd\u56e2\u961f\u7684\u8def\u5b50\uff0c\u5148\u91cd\u5efa\u4e09\u7ef4\u573a\u666f\uff0c\u518d\u5728\u573a\u666f\u56fe\u4e0a\u505a\u63a8\u7406\u3002<\/p>\n

\u7ed3\u679c\u53d1\u73b0\uff0c\u5982\u679c\u7ed9\u7684\u662f\u771f\u503c3D\uff08\u4e0a\u5e1d\u89c6\u89d2\u7684\u5b8c\u7f8e\u51e0\u4f55\uff09\uff0c\u90a3\u786e\u5b9e\u5f88\u5f3a\u3002<\/p>\n

Gemini\u5728\u6750\u8d28\u900f\u660e\u4efb\u52a1\u4e0a\uff0c2D\u7248\u672c\u5f97\u520644.0%\uff0c3D\u7248\u672c\u5f97\u520660.4%\uff0c\u63d0\u534716.4\u4e2a\u767e\u5206\u70b9\u3002\u5728\u9700\u8981\u7cbe\u786e\u6df1\u5ea6\u4fe1\u606f\u7684\u4efb\u52a1\u4e0a\uff0c3D grounding\u6709\u5929\u7136\u4f18\u52bf\u3002<\/p>\n

\u4f46\u5982\u679c\u662f\u771f\u5b9e\u91cd\u5efa\u5462\uff1f\u56e2\u961f\u7528\u5f53\u524d\u6700\u5148\u8fdb\u7684VGGT\u6a21\u578b\u505a\u573a\u666f\u91cd\u5efa\uff0c\u518d\u628a\u91cd\u5efa\u7ed3\u679c\u5582\u7ed9\u63a8\u7406\u6a21\u578b\u3002<\/p>\n

\u7ed3\u679c\u90a3\u53eb\u4e00\u4e2a\u60e8\u4e0d\u5fcd\u7779\uff1a\u51e0\u4f55\u914d\u7f6e\u4efb\u52a1\u4e0a\uff0c2D\u57fa\u7ebf\u5f97\u520627.5%\uff0cVGGT\u91cd\u5efa\u540e\u7684\u573a\u666f\u56fe\u5f97\u5206\u53ea\u67099.9%\u3002<\/p>\n

\u8fd9\u8bf4\u660e\uff0c\u4e0d\u5b8c\u7f8e\u76843D\u4e0d\u662f\u4e2d\u6027\u5931\u8d25\uff0c\u5b83\u662f\u8d1f\u5411\u5931\u8d25\u3002\u51e0\u4f55\u4f2a\u5f71\u3001\u906e\u6321\u8865\u5168\u9519\u8bef\u3001\u6df1\u5ea6\u4f30\u8ba1\u504f\u5dee\uff0c\u628a\u8fd9\u4e9b\u5931\u771f\u4fe1\u606f\u7f16\u7801\u6210\u573a\u666f\u56fe\uff0c\u5c31\u7b49\u4e8e\u7ed9\u63a8\u7406\u6a21\u578b\u5582\u4e86\u4e00\u4efd\u300c\u6709\u6bd2\u300d\u7684\u8f93\u5165\u3002<\/p>\n

\u76f8\u6bd4\u4e4b\u4e0b\uff0c2D\u867d\u7136\u4fe1\u606f\u5c11\uff0c\u4f46\u81f3\u5c11\u4e0d\u5931\u771f\uff1b3D\u5982\u679c\u91cd\u5efa\u8d28\u91cf\u4e0d\u8fc7\u5173\uff0c\u6bd42D\u8fd8\u4e0d\u5982\u3002<\/p>\n

\u7b2c\u4e09\uff0c\u5143\u8ba4\u77e5\u7f3a\u9677\uff1a\u6a21\u578b\u4e0d\u77e5\u9053\u81ea\u5df1\u770b\u6ca1\u770b\u591f\u3002<\/p>\n

\u8bba\u6587\u91cc\u8fd8\u6709\u4e00\u7ec4\u5bf9\u6bd4\u5b9e\u9a8c\uff0c\u63a2\u8ba8\u4e86\u667a\u80fd\u4f53\u548c\u4eba\u7c7b\u7684\u7a7a\u95f4\u63a8\u7406\u80fd\u529b\u7a76\u7adf\u8fd8\u6709\u591a\u5927\u5dee\u8ddd\u3002<\/p>\n

\u7ed3\u679c\u53d1\u73b0\uff0c\u5c3d\u7ba1\u4eba\u7c7b\u4e0e\u6a21\u578b\u4e4b\u95f4\u5b58\u5728\u611f\u77e5\u5dee\u8ddd\uff0c\u4f46\u8be5\u5dee\u8ddd\u53ef\u80fd\u6bd4\u666e\u904d\u8ba4\u4e3a\u7684\u8981\u5c0f\u3002<\/p>\n

\u5728\u90e8\u5206\u7c7b\u522b\u4e2d\uff0c\u6a21\u578b\u7684\u88ab\u52a8\u8868\u73b0\u751a\u81f3\u80fd\u4e0e\u4eba\u7c7b\u6301\u5e73\u6216\u8d85\u8d8a\u4eba\u7c7b\u3002<\/p>\n

\u5728\u771f\u5b9e\u8f68\u8ff9\u6761\u4ef6\u4e0b\uff0cGemini\u5728\u90e8\u5206\u906e\u6321\u4efb\u52a1\u4e0a\u8fbe\u523088.4%\u7684\u51c6\u786e\u7387\uff0c\u800c\u4eba\u7c7b\u4e3a87.4%\uff1bGPT-5\u5728\u6750\u8d28\u900f\u660e\u5ea6\u4efb\u52a1\u4e0a\u8fbe\u523096.3%\uff0c\u4eba\u7c7b\u5219\u4e3a97.2%\u3002<\/p>\n

\u7136\u800c\u5728\u4e3b\u52a8\u63a2\u7d22\u573a\u666f\u4e0b\uff0c\u4e8c\u8005\u7684\u5dee\u8ddd\u6025\u5267\u663e\u73b0\u3002<\/p>\n

\u4eba\u7c7b\u51ed\u501f\u660e\u786e\u7684\u89c2\u5bdf\u76ee\u6807\u548c\u505c\u6b62\u65f6\u673a\uff0c\u8868\u73b0\u8fdc\u8d85\u6a21\u578b\uff0c\u4e14\u4e3b\u52a8\u63a2\u7d22\u7684\u8868\u73b0\u66f4\u63a5\u8fd1\u771f\u5b9e\u8f68\u8ff9\u4e0b\u7684\u88ab\u52a8\u8868\u73b0\u3002<\/p>\n

\u4f8b\u5982\u5728\u7269\u7406\u63a5\u89e6\u4efb\u52a1\u4e2d\uff0c\u4eba\u7c7b\u51c6\u786e\u7387\u4e3a88.3%\uff0c\u800c GPT-5\u4ec5\u4e3a 64.2%\uff1b\u5728\u6750\u8d28\u900f\u660e\u5ea6\u4efb\u52a1\u4e2d\uff0c\u4eba\u7c7b\u51c6\u786e\u7387\u4e3a93.6%\uff0cGemini 3.1\u5219\u4e3a52.3%\u3002<\/p>\n

\u901a\u8fc7\u5206\u6790\u6a21\u578b\u4e0e\u4eba\u7c7b\u7684\u63a2\u7d22\u8f68\u8ff9\uff0c\u56e2\u961f\u53d1\u73b0\u4eba\u7c7b\u8868\u73b0\u51fa\u66f4\u5f3a\u7684\u8ba4\u77e5\u8c28\u614e\u6027\uff1a\u5728\u505a\u51fa\u5224\u65ad\u524d\u4f1a\u6536\u96c6\u66f4\u591a\u89c2\u6d4b\uff0c\u4e3b\u52a8\u5bfb\u627e\u53ef\u80fd\u8bc1\u4f2a\u5f53\u524d\u5047\u8bbe\u7684\u89c6\u89d2\uff0c\u5e76\u5728\u6a21\u7cca\u60c5\u5883\u4e0b\u964d\u4f4e\u7f6e\u4fe1\u5ea6\u3002<\/p>\n

\u800c\u6a21\u578b\u5219\u4f1a\u8fc7\u65e9\u505c\u6b62\u63a2\u7d22\uff0c\u5373\u4fbf\u8bc1\u636e\u5b58\u5728\u6a21\u7cca\u6027\uff0c\u4e5f\u4ec5\u5728\u5c11\u6570\u6b65\u9aa4\u540e\u5c31\u4ee5\u9ad8\u7f6e\u4fe1\u5ea6\u505a\u51fa\u5224\u65ad\uff0c\u8fdb\u800c\u4ea7\u751f\u4e0e\u573a\u666f\u72b6\u6001\u76f8\u6096\u7684\u7a7a\u95f4\u5e7b\u89c9\u3002<\/p>\n

\u6a21\u578b\u7684\u8fc7\u5ea6\u81ea\u4fe1\uff0c\u8fd8\u56e0\u52a8\u4f5c\u9009\u62e9\u7684\u65b9\u5411\u504f\u5dee\u800c\u52a0\u5267\uff1a\u6a21\u578b\u4e0d\u4f1a\u63a2\u67e5\u6b63\u4ea4\u89d2\u5ea6\u6216\u5bfb\u627e\u80fd\u63a8\u7ffb\u521d\u59cb\u5370\u8c61\u7684\u89c6\u89d2\uff0c\u800c\u662f\u53cd\u590d\u5411\u540c\u4e00\u65b9\u5411\u79fb\u52a8\uff0c\u79ef\u7d2f\u7684\u662f\u5197\u4f59\u4fe1\u606f\u800c\u975e\u6709\u6548\u89c2\u6d4b\u3002<\/p>\n

\u56e2\u961f\u628a\u5b83\u5b9a\u6027\u4e3a\u5143\u8ba4\u77e5\uff08metacognition\uff09\u7f3a\u9677\uff1a\u6a21\u578b\u4e0d\u77e5\u9053\u81ea\u5df1\u4e0d\u77e5\u9053\u3002<\/p>\n

\u5b83\u7f3a\u4e4f\u4e00\u79cd\u5185\u5efa\u7684\u300c\u6000\u7591\u673a\u5236\u300d\uff0c\u65e0\u6cd5\u8bc4\u4f30\u5f53\u524d\u4fe1\u606f\u662f\u5426\u5145\u5206\uff0c\u65e0\u6cd5\u6839\u636e\u77db\u76fe\u8bc1\u636e\u8c03\u6574\u4fe1\u5ff5\u3002<\/p>\n

\u8fd9\u4e2a\u95ee\u9898\u4ece\u6839\u672c\u4e0a\u533a\u522b\u4e8e\u611f\u77e5\u80fd\u529b\uff0c\u4e5f\u662f\u4e00\u4e2a\u66f4\u52a0\u5e95\u5c42\u7684\u6311\u6218\uff0c\u4ec5\u9760\u66f4\u5f3a\u7684\u89c6\u89c9\u7f16\u7801\u5668\u6216\u66f4\u591a\u7684\u63a2\u7d22\u6b65\u9aa4\u65e0\u6cd5\u89e3\u51b3\u3002<\/p>\n

\u6700\u540e\uff0c\u518d\u4ecb\u7ecd\u4e00\u4e0b\u8fd9\u9879\u5de5\u4f5c\u7684\u4f5c\u8005\u56e2\u961f\u3002<\/p>\n

\u4e00\u4f5c\u662fYining Hong\u3002<\/p>\n

Yining Hong\uff0c\u65af\u5766\u798f\u5927\u5b66\u7684\u535a\u58eb\u540e\uff0c\u5bfc\u5e08\u4e3aYejin Choi\u6559\u6388\uff0c\u540c\u65f6\u53d7\u5230Leonidas Guibas\u6559\u6388\u3001\u5434\u5bb6\u4fca\u6559\u6388\u548c\u674e\u98de\u98de\u6559\u6388\u7684\u5bc6\u5207\u6307\u5bfc\u3002<\/p>\n

\u5979\u66fe\u5728UCLA\u83b7\u5f97\u8ba1\u7b97\u673a\u79d1\u5b66\u535a\u58eb\u5b66\u4f4d\uff0c\u672c\u79d1\u5c31\u8bfb\u4e8e\u4e0a\u6d77\u4ea4\u901a\u5927\u5b66\u7535\u5b50\u5de5\u7a0b\u7cfb\u3002<\/p>\n

\u6b64\u5916\uff0c\u5979\u8fd8\u662f\u4e00\u540d\u804c\u4e1a\u97f3\u4e50\u5bb6\uff0c\u5e73\u65f6\u4f1a\u548c\u4e50\u961f\u4e00\u8d77\u5de1\u6f14\uff0c\u540c\u65f6\u4e5f\u662fCVPR 2026\u7684\u793e\u4ea4\u4e3b\u5e2d\uff0c\u8d1f\u8d23\u7ec4\u7ec7CVPR\u62db\u5f85\u4f1a\u548c\u97f3\u4e50\u8868\u6f14\u3002<\/p>\n

Jiageng Liu\uff08\u5218\u5bb6\u8015\uff09\uff0c\u52a0\u5dde\u5927\u5b66\u6d1b\u6749\u77f6\u5206\u6821\uff08UCLA\uff09Mobility Lab\u7684\u535a\u58eb\u751f\u3002<\/p>\n

\u5176\u672c\u79d1\u5c31\u8bfb\u4e8e\u6d59\u6c5f\u5927\u5b66\u7afa\u53ef\u6862\u8363\u8a89\u5b66\u9662\u53ca\u8ba1\u7b97\u673a\u79d1\u5b66\u4e0e\u6280\u672f\u5b66\u9662\u7684\u56fe\u7075\u73ed\uff0c\u83b7\u4eba\u5de5\u667a\u80fd\u5b66\u58eb\u5b66\u4f4d\u3002<\/p>\n

Han Yin\uff0c\u6e05\u534e\u5927\u5b66\u672c\u79d1\u751f\uff0c\u65af\u5766\u798f\u5927\u5b66Intern\uff0c\u4e13\u4e1a\u4e3a\u8ba1\u7b97\u673a\u79d1\u5b66\u4e0e\u6280\u672f\u3002<\/p>\n

\u674e\u98de\u98de\u3001\u5434\u4f73\u4fca\uff08Jiajun Wu\uff09\u3001Yejin Choi\uff0c\u4e09\u4f4d\u65af\u5766\u798f\u6559\u6388\uff0c\u4e5f\u540c\u65f6\u51fa\u73b0\u5728\u4f5c\u8005\u5217\u8868\u91cc\u3002<\/p>\n

\u53e6\u5916\u8fd8\u6709\u6765\u81ea\u897f\u5317\u5927\u5b66\u7684Manling Li\u6559\u6388\u548c\u65af\u5766\u798f\u7684Leonidas Guibas\u6559\u6388\u53c2\u4e0e\u3002<\/p>\n

\u53c2\u8003\u94fe\u63a5\uff1a[1]https:\/\/arxiv.org\/abs\/2605.18746[2]https:\/\/esi-bench.github.io\/<\/p>\n

\n

\u7248\u6743\u58f0\u660e\uff1a<\/strong><\/p>\n

\u672c\u535a\u5ba2\u90e8\u5206\u5185\u5bb9\u4e3a\u8f6c\u8f7d\u6587\u7ae0\uff0c\u65e8\u5728\u5206\u4eab\u6709\u4ef7\u503c\u7684\u4fe1\u606f\uff0c\u7248\u6743\u5f52\u539f\u4f5c\u8005\u6240\u6709\u3002<\/p>\n

\u8f6c\u8f7d\u4ec5\u4e3a\u4e2a\u4eba\u5b66\u4e60\u4e0e\u4ea4\u6d41\u76ee\u7684\uff0c\u4e0d\u5bf9\u6587\u7ae0\u89c2\u70b9\u8d1f\u8d23\uff0c\u4ea6\u4e0d\u7528\u4e8e\u4efb\u4f55\u5546\u4e1a\u7528\u9014\u3002<\/p>\n

\u5982\u6d89\u53ca\u7248\u6743\u95ee\u9898\uff0c\u8bf7\u8054\u7cfb\u672c\u4eba\u5220\u9664\u3002<\/p>\n<\/blockquote>\n","protected":false},"excerpt":{"rendered":"

\u4e00\u4e2a\u4e13\u95e8\u7528\u6765\u8bc4\u6d4b\u5177\u8eab\u7a7a\u95f4\u667a\u80fd\u7684\u65b0\u57fa\u51c6<\/p>\n

\u542c\u96e8 \u53d1\u81ea \u51f9\u975e\u5bfa<\/p>\n

ImageNet\u4e4b\u540e\uff0c\u674e\u98de\u98de\u518d\u51fa\u624b\uff01<\/p>\n

\u674e\u98de\u98de\u56e2\u961f\u6700\u65b0\u53d1\u5e03ESI-Bench\u2014\u2014\u4e00\u4e2a\u4e13\u95e8\u7528\u6765\u8bc4\u6d4b\u5177\u8eab\u7a7a\u95f4\u667a\u80fd\u7684\u65b0\u57fa\u51c6\u3002<\/p>\n

\u8fc7\u53bb\u7684\u7a7a\u95f4\u667a\u80fd\u8bc4\u6d4b\u9ed8\u8ba4\u7ed9\u6a21\u578b\u6700\u4f18\u89c2\u6d4b\uff0c\u800cESI-Bench\u7b2c\u4e00\u4e2a\u628a\u89c2\u5bdf\u8005\u53d8\u6210\u884c\u52a8\u8005\uff0c\u95ed\u5408\u4e86\u611f\u77e5-\u884c\u52a8\u56de\u8def\u3002<\/p>\n

\u5b83\u4e3a\u5177\u8eab\u7a7a\u95f4\u667a\u80fd\u9886\u57df\u63d0\u4f9b\u4e86\u4e00\u4e2a\u7cfb\u7edf\u6027\u7684\u8bc4\u6d4b\u6846\u67b6\uff0c\u8986\u76d6\u4eba\u7c7b\u6838\u5fc3\u7a7a\u95f4\u8ba4\u77e5\u80fd\u529b\u7684\u56db\u5927\u7ef4\u5ea6\u3002<\/p>\n

\u8bba\u6587\u7684\u6838\u5fc3\u7ed3\u8bba\u662f\uff1a<\/p>\n","protected":false},"author":1,"featured_media":5122,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[2,295],"class_list":["post-5126","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-xwzx","tag-ai","tag-295"],"_links":{"self":[{"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/posts\/5126","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/comments?post=5126"}],"version-history":[{"count":0,"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/posts\/5126\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/media\/5122"}],"wp:attachment":[{"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/media?parent=5126"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/categories?post=5126"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/1ow.top\/index.php\/wp-json\/wp\/v2\/tags?post=5126"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}