{"id":5639,"date":"2025-02-17T11:46:30","date_gmt":"2025-02-17T03:46:30","guid":{"rendered":"https:\/\/nullthought.net\/?p=5639"},"modified":"2025-02-23T11:08:04","modified_gmt":"2025-02-23T03:08:04","slug":"embodiedbench%ef%bc%9a%e4%b8%93%e9%97%a8%e7%94%a8%e4%ba%8e%e8%af%84%e4%bc%b0%e8%a7%86%e8%a7%89%e9%a9%b1%e5%8a%a8%e5%85%b7%e8%ba%ab%e6%99%ba%e8%83%bd%e4%bd%93%ef%bc%88vision-driven-embodied-agents","status":"publish","type":"post","link":"https:\/\/nullthought.net\/?p=5639","title":{"rendered":"EmbodiedBench\uff1a\u4e13\u95e8\u7528\u4e8e\u8bc4\u4f30\u89c6\u89c9\u9a71\u52a8\u5177\u8eab\u667a\u80fd\u4f53\uff08Vision-Driven Embodied Agents\uff09\u7684\u57fa\u51c6\u6d4b\u8bd5\u6846\u67b6"},"content":{"rendered":"\n<p>\u8fd1\u5e74\u6765\uff0c\u5177\u8eab\u667a\u80fd\u4f53\uff08Embodied Agents\uff09\u5728\u4eba\u5de5\u667a\u80fd\u7814\u7a76\u4e2d\u53d7\u5230\u4e86\u8d8a\u6765\u8d8a\u591a\u7684\u5173\u6ce8\u3002\u8fd9\u4e9b\u667a\u80fd\u4f53\u9700\u8981\u5177\u5907\u7406\u89e3\u73af\u5883\u3001\u611f\u77e5\u89c6\u89c9\u4fe1\u606f\u4ee5\u53ca\u6267\u884c\u590d\u6742\u4efb\u52a1\u7684\u80fd\u529b\u3002\u591a\u6a21\u6001\u5927\u578b\u8bed\u8a00\u6a21\u578b\uff08Multi-modal Large Language Models, MLLMs\uff09\u7684\u5174\u8d77\uff0c\u4f7f\u5f97\u57fa\u4e8e\u5927\u89c4\u6a21\u9884\u8bad\u7ec3\u6a21\u578b\u7684\u667a\u80fd\u4f53\u6210\u4e3a\u53ef\u80fd\u3002\u7136\u800c\uff0c\u5f53\u524d\u7814\u7a76\u4e3b\u8981\u96c6\u4e2d\u5728\u8bed\u8a00\u9a71\u52a8\u7684\u5177\u8eab\u667a\u80fd\u4f53\uff0c\u800c MLLMs \u5728\u5177\u8eab\u667a\u80fd\u4f53\u4e2d\u7684\u5e94\u7528\u4ecd\u7136\u7f3a\u4e4f\u7cfb\u7edf\u6027\u7684\u8bc4\u4f30\u6846\u67b6\u3002<\/p>\n\n\n\n<p>\u8bba\u6587<strong><a href=\"https:\/\/arxiv.org\/abs\/2502.09560\" target=\"_blank\" rel=\"noreferrer noopener\">EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents<\/a><\/strong>\u63d0\u51fa <strong>EmbodiedBench<\/strong>\uff0c\u8fd9\u662f\u4e00\u4e2a\u5168\u9762\u7684\u57fa\u51c6\u6d4b\u8bd5\u6846\u67b6\uff0c\u4e13\u95e8\u7528\u4e8e\u8bc4\u4f30<strong>\u89c6\u89c9\u9a71\u52a8\u7684\u5177\u8eab\u667a\u80fd\u4f53<\/strong>\u3002\u5176\u4e3b\u8981\u7279\u70b9\u5305\u62ec\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u591a\u5c42\u6b21\u4efb\u52a1\u8bc4\u4f30<\/strong>\uff1a\u6db5\u76d6 <strong>1,128 \u4e2a\u6d4b\u8bd5\u4efb\u52a1<\/strong>\uff0c\u6d89\u53ca<strong>\u56db\u4e2a\u4e0d\u540c\u73af\u5883<\/strong>\uff0c\u5305\u62ec\u9ad8\u5c42\u8bed\u4e49\u4efb\u52a1\uff08\u5982\u5bb6\u5ead\u573a\u666f\uff09\u548c\u4f4e\u5c42\u7269\u7406\u4ea4\u4e92\u4efb\u52a1\uff08\u5982\u5bfc\u822a\u548c\u64cd\u63a7\uff09\u3002<\/li>\n\n\n\n<li><strong>\u7ec6\u7c92\u5ea6\u80fd\u529b\u6d4b\u8bc4<\/strong>\uff1a\u8bbe\u8ba1\u516d\u4e2a\u5b50\u96c6\uff0c\u4e13\u95e8\u8bc4\u4f30\u667a\u80fd\u4f53\u7684<strong>\u5e38\u8bc6\u63a8\u7406\u3001\u590d\u6742\u6307\u4ee4\u7406\u89e3\u3001\u7a7a\u95f4\u611f\u77e5\u3001\u89c6\u89c9\u611f\u77e5\u3001\u957f\u65f6\u89c4\u5212<\/strong>\u7b49\u5173\u952e\u80fd\u529b\u3002<\/li>\n<\/ol>\n\n\n\n<p>\u901a\u8fc7\u5bf9 13 \u4e2a\u4e3b\u6d41 MLLMs\uff08\u5305\u62ec\u4e13\u6709\u548c\u5f00\u6e90\u6a21\u578b\uff09\u8fdb\u884c\u6d4b\u8bd5\uff0c\u7814\u7a76\u53d1\u73b0\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u73b0\u6709 MLLMs \u5728\u9ad8\u5c42\u6b21\u4efb\u52a1\uff08\u5982\u8ba1\u5212\u4efb\u52a1\uff09\u4e0a\u8868\u73b0\u8f83\u597d\uff0c\u4f46\u5728\u4f4e\u5c42\u7269\u7406\u64cd\u63a7\u4efb\u52a1\u4e0a\u5b58\u5728\u663e\u8457\u6027\u80fd\u74f6\u9888\uff08\u6700\u4f73\u6a21\u578b GPT-4o \u7684\u5e73\u5747\u6210\u529f\u7387\u4ec5\u4e3a <strong>28.9%<\/strong>\uff09\u3002<\/li>\n\n\n\n<li>\u89c6\u89c9\u8f93\u5165\u5bf9\u4e8e\u4f4e\u5c42\u6b21\u4efb\u52a1\u81f3\u5173\u91cd\u8981\uff0c\u5176\u53bb\u9664\u540e\u6210\u529f\u7387\u4e0b\u964d <strong>40%-70%<\/strong>\uff0c\u800c\u9ad8\u5c42\u4efb\u52a1\u5bf9\u89c6\u89c9\u8f93\u5165\u7684\u4f9d\u8d56\u8f83\u5c0f\u3002<\/li>\n<\/ul>\n\n\n\n<p>\u672c\u8bba\u6587\u4f5c\u4e3aRui Yang, Hanyang Chen, Junyu Zhang, Mark Zhao, Cheng Qian, Kangrui Wang, Qineng Wang, Teja Venkat Koripella, Marziyeh Movahedi, Manling Li, Heng Ji, Huan Zhang, Tong Zhang\uff0c\u6765\u81eaUniversity of Illinois Urbana-Champaign, Northwestern University, University of Toronto\u548cToyota Technological Institute at Chicago\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"560\" src=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/02\/image-8-1024x560.png\" alt=\"\" class=\"wp-image-5640\" srcset=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/02\/image-8-1024x560.png 1024w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/02\/image-8-300x164.png 300w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/02\/image-8-768x420.png 768w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/02\/image-8.png 1041w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><figcaption class=\"wp-element-caption\"><strong><a href=\"https:\/\/arxiv.org\/abs\/2502.09560\" target=\"_blank\" rel=\"noreferrer noopener\">EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents<\/a><\/strong><\/figcaption><\/figure>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e00\u3001\u76f8\u5173\u7814\u7a76<\/h5>\n\n\n\n<p>\u5177\u8eab\u667a\u80fd\u4f53\u7684\u7814\u7a76\u53ef\u4ee5\u8ffd\u6eaf\u5230\u65e9\u671f\u7684\u5f3a\u5316\u5b66\u4e60\u548c\u673a\u5668\u4eba\u7814\u7a76\uff0c\u8fd1\u5e74\u6765\u968f\u7740 LLMs \u548c MLLMs \u7684\u53d1\u5c55\uff0c\u8fd9\u4e9b\u6a21\u578b\u5728\u63a8\u7406\u548c\u51b3\u7b56\u4efb\u52a1\u4e0a\u5c55\u73b0\u4e86\u5f3a\u5927\u7684\u80fd\u529b\u3002\u7136\u800c\uff0c\u76ee\u524d\u7684\u8bc4\u6d4b\u6846\u67b6\u4ecd\u7136\u5b58\u5728\u5c40\u9650\u6027\u3002<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>LLMs \u5728\u9ad8\u5c42\u4efb\u52a1\u4e2d\u7684\u5e94\u7528<\/strong>\n<ul class=\"wp-block-list\">\n<li>SayCan\uff08Ahn et al., 2022\uff09\uff1a\u7ed3\u5408\u8bed\u8a00\u6a21\u578b\u548c\u5f3a\u5316\u5b66\u4e60\uff0c\u4f7f\u673a\u5668\u4eba\u80fd\u6839\u636e\u81ea\u7136\u8bed\u8a00\u6307\u4ee4\u6267\u884c\u4efb\u52a1\u3002<\/li>\n\n\n\n<li>Inner Monologue\uff08Huang et al., 2022\uff09\uff1a\u5f15\u5165\u53cd\u601d\u673a\u5236\uff0c\u8ba9\u667a\u80fd\u4f53\u5728\u6267\u884c\u8fc7\u7a0b\u4e2d\u8c03\u6574\u51b3\u7b56\u3002<\/li>\n\n\n\n<li>VOXPOSER\uff08Huang et al., 2023\uff09\uff1a\u901a\u8fc7 3D \u7ed3\u6784\u5efa\u6a21\u63d0\u5347\u673a\u5668\u4eba\u5728\u590d\u6742\u573a\u666f\u4e2d\u7684\u9002\u5e94\u80fd\u529b\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>MLLMs \u5728\u611f\u77e5\u548c\u51b3\u7b56\u4e2d\u7684\u5e94\u7528<\/strong>\n<ul class=\"wp-block-list\">\n<li>PALM-E\uff08Driess et al., 2023\uff09\uff1a\u4e00\u4e2a\u7aef\u5230\u7aef\u7684\u591a\u6a21\u6001\u6a21\u578b\uff0c\u7ed3\u5408\u89c6\u89c9\u548c\u8bed\u8a00\u4fe1\u606f\uff0c\u5b9e\u73b0\u8de8\u9886\u57df\u4efb\u52a1\u3002<\/li>\n\n\n\n<li>RT-2\uff08Brohan et al., 2023\uff09\uff1a\u91c7\u7528 Transformer \u7ed3\u6784\uff0c\u5c06\u5927\u89c4\u6a21\u7f51\u7edc\u77e5\u8bc6\u8fc1\u79fb\u5230\u673a\u5668\u4eba\u63a7\u5236\u9886\u57df\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u73b0\u6709\u57fa\u51c6\u6d4b\u8bd5\u7684\u5c40\u9650\u6027<\/strong>\n<ul class=\"wp-block-list\">\n<li><strong>ALFWorld<\/strong>\u3001<strong>VisualAgentBench<\/strong>\u3001<strong>AgentBench<\/strong> \u7b49\u8bc4\u6d4b\u6846\u67b6\u4e3b\u8981\u9488\u5bf9 LLMs\uff0c\u7f3a\u4e4f\u5bf9 MLLMs \u5728\u5177\u8eab\u4efb\u52a1\u4e2d\u7684\u8bc4\u6d4b\u3002<\/li>\n\n\n\n<li><strong>VisualAgentBench<\/strong> \u662f\u76ee\u524d\u552f\u4e00\u9488\u5bf9 MLLMs \u8bc4\u6d4b\u7684\u6846\u67b6\uff0c\u4f46\u5b83\u4e3b\u8981\u5173\u6ce8\u9ad8\u5c42\u4efb\u52a1\uff0c\u5982\u5bb6\u5ead\u4efb\u52a1\u6216 Minecraft \u6e38\u620f\uff0c\u800c\u4f4e\u5c42\u4efb\u52a1\uff08\u5982\u7cbe\u51c6\u64cd\u63a7\u548c\u7a7a\u95f4\u611f\u77e5\uff09\u672a\u5f97\u5230\u5145\u5206\u7814\u7a76\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<p>\u4e3a\u5f25\u8865\u8fd9\u4e9b\u4e0d\u8db3\uff0c<strong>EmbodiedBench \u8bbe\u8ba1\u4e86\u591a\u5c42\u6b21\u4efb\u52a1\u548c\u80fd\u529b\u8bc4\u6d4b\uff0c\u6db5\u76d6\u4ece\u9ad8\u5c42\u8ba1\u5212\u5230\u4f4e\u5c42\u64cd\u63a7\u7684\u5b8c\u6574\u4efb\u52a1\u4f53\u7cfb<\/strong>\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e8c\u3001\u95ee\u9898\u5efa\u6a21<\/h5>\n\n\n\n<p>\u5177\u8eab\u667a\u80fd\u4f53\u7684\u884c\u4e3a\u53ef\u4ee5\u5206\u4e3a\u9ad8\u5c42\u4efb\u52a1\u548c\u4f4e\u5c42\u4efb\u52a1\uff0c\u6bcf\u79cd\u4efb\u52a1\u7684\u590d\u6742\u6027\u548c\u6267\u884c\u65b9\u5f0f\u4e0d\u540c\u3002<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u5177\u8eab\u667a\u80fd\u4f53\u7684\u884c\u52a8\u5c42\u6b21<\/strong>\n<ul class=\"wp-block-list\">\n<li><strong>\u4f4e\u5c42\u6b21\u884c\u52a8\uff08Low-Level Actions\uff09<\/strong>\uff1a\u6307\u76f4\u63a5\u53ef\u6267\u884c\u7684\u539f\u5b50\u7ea7\u6307\u4ee4\uff0c\u5982 <code>move forward 0.1m<\/code>\u3002\u901a\u5e38\u91c7\u7528 7 \u7ef4\u5411\u91cf <code>[X, Y, Z, Roll, Pitch, Yaw, Gripper]<\/code> \u8fdb\u884c\u63cf\u8ff0\u3002<\/li>\n\n\n\n<li><strong>\u9ad8\u5c42\u6b21\u884c\u52a8\uff08High-Level Actions\uff09<\/strong>\uff1a\u7531\u591a\u4e2a\u4f4e\u5c42\u884c\u52a8\u7ec4\u6210\uff0c\u4f8b\u5982 <code>find a HandTowel<\/code> \u53ef\u80fd\u6d89\u53ca\u65cb\u8f6c\u3001\u626b\u63cf\u76ee\u6807\u3001\u63a5\u8fd1\u76ee\u6807\u7b49\u591a\u4e2a\u5b50\u6b65\u9aa4\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u89c6\u89c9\u9a71\u52a8\u667a\u80fd\u4f53\u7684\u51b3\u7b56\u5efa\u6a21<\/strong><br>\u667a\u80fd\u4f53\u7684\u51b3\u7b56\u53ef\u5efa\u6a21\u4e3a<strong>\u90e8\u5206\u53ef\u89c2\u6d4b\u9a6c\u5c14\u53ef\u592b\u51b3\u7b56\u8fc7\u7a0b\uff08POMDP\uff09<\/strong>\uff1a<\/li>\n<\/ol>\n\n\n\n<h4 class=\"wp-block-heading\">(S,A,\u03a9,T,O,L,R)<\/h4>\n\n\n\n<p>\u5176\u4e2d\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>S\uff1a\u5b8c\u6574\u73af\u5883\u72b6\u6001\uff08\u4e0d\u53ef\u76f4\u63a5\u89c2\u6d4b\uff09<\/li>\n\n\n\n<li>A\uff1a\u9ad8\u5c42\u6216\u4f4e\u5c42\u884c\u52a8\u7a7a\u95f4<\/li>\n\n\n\n<li>\u03a9\uff1a\u89c6\u89c9\u611f\u77e5\u7a7a\u95f4<\/li>\n\n\n\n<li>T\uff1a\u72b6\u6001\u8f6c\u79fb\u51fd\u6570<\/li>\n\n\n\n<li>O\uff1a\u72b6\u6001\u4e0e\u89c2\u6d4b\u4e4b\u95f4\u7684\u6620\u5c04<\/li>\n\n\n\n<li>L\uff1a\u8bed\u8a00\u6307\u4ee4<\/li>\n\n\n\n<li>R\uff1a\u4efb\u52a1\u5b8c\u6210\u7684\u5956\u52b1\u51fd\u6570<\/li>\n<\/ul>\n\n\n\n<p>\u667a\u80fd\u4f53\u7684\u76ee\u6807\u662f\u901a\u8fc7\u7b56\u7565\u03c0(a<sub>t<\/sub>\u200b\u2223L,h<sub>t<\/sub>\u200b) <strong>\u6700\u5927\u5316\u4efb\u52a1\u6210\u529f\u7387<\/strong>\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e09\u3001EmbodiedBench \u8bbe\u8ba1<\/h5>\n\n\n\n<p>EmbodiedBench \u5305\u542b\u56db\u4e2a\u4e0d\u540c\u73af\u5883\uff0c\u6bcf\u4e2a\u73af\u5883\u9488\u5bf9\u7279\u5b9a\u4efb\u52a1\u8fdb\u884c\u8bc4\u6d4b\u3002<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>EB-ALFRED<\/strong>\uff08\u57fa\u4e8e ALFRED \u6570\u636e\u96c6\uff09\uff1a\n<ul class=\"wp-block-list\">\n<li>\u4efb\u52a1\uff1a<strong>\u9ad8\u5c42\u89c4\u5212<\/strong>\uff0c\u6d89\u53ca<strong>\u62fe\u53d6\u3001\u653e\u7f6e\u3001\u6253\u5f00\u3001\u5173\u95ed\u7b49\u64cd\u4f5c<\/strong><\/li>\n\n\n\n<li>\u8bc4\u6d4b\u80fd\u529b\uff1a\u4efb\u52a1\u5206\u89e3\u3001\u957f\u671f\u8ba1\u5212<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>EB-Habitat<\/strong>\uff08\u57fa\u4e8e Language Rearrangement Benchmark\uff09\uff1a\n<ul class=\"wp-block-list\">\n<li>\u4efb\u52a1\uff1a<strong>\u5bfc\u822a\u548c\u7269\u54c1\u653e\u7f6e<\/strong><\/li>\n\n\n\n<li>\u8bc4\u6d4b\u80fd\u529b\uff1a\u76ee\u6807\u5bfc\u5411\u4efb\u52a1\u6267\u884c\u3001\u5c40\u90e8\u73af\u5883\u63a2\u7d22<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>EB-Navigation<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u4efb\u52a1\uff1a<strong>\u76ee\u6807\u5bfc\u5411\u5bfc\u822a<\/strong><\/li>\n\n\n\n<li>\u8bc4\u6d4b\u80fd\u529b\uff1a\u7a7a\u95f4\u611f\u77e5\u3001\u8def\u5f84\u89c4\u5212<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>EB-Manipulation<\/strong>\uff08\u57fa\u4e8e VLMBench\uff09\uff1a\n<ul class=\"wp-block-list\">\n<li>\u4efb\u52a1\uff1a<strong>\u4f4e\u5c42\u6b21\u7269\u4f53\u64cd\u63a7<\/strong><\/li>\n\n\n\n<li>\u8bc4\u6d4b\u80fd\u529b\uff1a\u7cbe\u51c6\u52a8\u4f5c\u6267\u884c\u3001\u7269\u4f53\u4ea4\u4e92<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<p>EmbodiedBench \u8bbe\u8ba1\u516d\u4e2a\u5b50\u96c6\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u57fa\u672c\u4efb\u52a1\u80fd\u529b\uff08Base\uff09<\/strong><\/li>\n\n\n\n<li><strong>\u5e38\u8bc6\u63a8\u7406\uff08Common Sense\uff09<\/strong><\/li>\n\n\n\n<li><strong>\u590d\u6742\u6307\u4ee4\u7406\u89e3\uff08Complex Instruction\uff09<\/strong><\/li>\n\n\n\n<li><strong>\u7a7a\u95f4\u611f\u77e5\uff08Spatial Awareness\uff09<\/strong><\/li>\n\n\n\n<li><strong>\u89c6\u89c9\u611f\u77e5\uff08Visual Perception\uff09<\/strong><\/li>\n\n\n\n<li><strong>\u957f\u65f6\u89c4\u5212\uff08Long Horizon\uff09<\/strong><\/li>\n<\/ol>\n\n\n\n<h5 class=\"wp-block-heading\">\u56db\u3001\u5b9e\u9a8c\u4e0e\u5206\u6790<\/h5>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u4e3b\u8981\u5b9e\u9a8c\u7ed3\u679c<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u9ad8\u5c42\u4efb\u52a1\u8868\u73b0\u4f18\u5f02\uff0c\u4f46\u4f4e\u5c42\u4efb\u52a1\u8868\u73b0\u4e0d\u4f73\uff0c\u5982 GPT-4o \u5728 EB-ALFRED \u53d6\u5f97 <strong>56.3%<\/strong> \u6210\u529f\u7387\uff0c\u800c\u5728 EB-Manipulation \u4ec5\u4e3a <strong>28.9%<\/strong>\u3002<\/li>\n\n\n\n<li>\u89c6\u89c9\u8f93\u5165\u5bf9\u4f4e\u5c42\u4efb\u52a1\u81f3\u5173\u91cd\u8981\uff0c\u53bb\u9664\u540e\u6210\u529f\u7387\u4e0b\u964d <strong>40%-70%<\/strong>\u3002<\/li>\n\n\n\n<li>\u957f\u65f6\u89c4\u5212\uff08Long-Horizon\uff09\u4e3a\u6700\u96be\u4efb\u52a1\uff0cGPT-4o \u5728\u6b64\u5b50\u96c6\u4ec5\u6709 <strong>64%<\/strong> \u6210\u529f\u7387\uff08\u76f8\u6bd4\u57fa\u7840\u4efb\u52a1\u4e0b\u964d 22%\uff09\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u8bef\u5dee\u5206\u6790<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u4e3b\u8981\u9519\u8bef\u7c7b\u578b\uff1a\n<ul class=\"wp-block-list\">\n<li>\u89c4\u5212\u9519\u8bef\uff0855%\uff09\uff1a\u5982\u7f3a\u5c11\u5fc5\u8981\u6b65\u9aa4<\/li>\n\n\n\n<li>\u8ba4\u77e5\u9519\u8bef\uff0841%\uff09\uff1a\u5982\u9519\u8bef\u7ec8\u6b62\u4efb\u52a1<\/li>\n\n\n\n<li>\u611f\u77e5\u9519\u8bef\uff084%\uff09\uff1a\u5982\u7269\u4f53\u8bc6\u522b\u9519\u8bef<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e94\u3001\u7ed3\u8bba\u4e0e\u672a\u6765\u65b9\u5411<\/h5>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u6539\u8fdb\u4f4e\u5c42\u4efb\u52a1\u6267\u884c<\/strong><\/li>\n\n\n\n<li><strong>\u589e\u5f3a\u957f\u65f6\u89c4\u5212\u80fd\u529b<\/strong><\/li>\n\n\n\n<li><strong>\u63d0\u5347\u89c6\u89c9\u611f\u77e5\u4e0e\u7a7a\u95f4\u63a8\u7406\u80fd\u529b<\/strong><\/li>\n\n\n\n<li><strong>\u4f18\u5316\u591a\u6b65\u89c6\u89c9\u7406\u89e3<\/strong><\/li>\n<\/ol>\n\n\n\n<p>EmbodiedBench \u4e3a\u672a\u6765 MLLM \u5728\u5177\u8eab\u667a\u80fd\u4f53\u4e2d\u7684\u5e94\u7528\u63d0\u4f9b\u4e86\u575a\u5b9e\u7684\u8bc4\u4f30\u5de5\u5177\uff0c\u5e76\u4e3a\u7814\u7a76\u8005\u63d0\u4f9b\u4e86\u66f4\u4e30\u5bcc\u7684\u5206\u6790\u65b9\u6cd5\u548c\u6539\u8fdb\u65b9\u5411\u3002<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p>EmbodiedBench on GitHub: <a href=\"https:\/\/embodiedbench.github.io\/\" target=\"_blank\" rel=\"noreferrer noopener\">https:\/\/embodiedbench.github.io\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u8fd1\u5e74\u6765\uff0c\u5177\u8eab\u667a\u80fd\u4f53\uff08Embodied Agents\uff09\u5728\u4eba\u5de5\u667a\u80fd\u7814\u7a76\u4e2d\u53d7\u5230\u4e86\u8d8a\u6765\u8d8a\u591a\u7684\u5173\u6ce8\u3002\u8fd9\u4e9b\u667a\u80fd\u4f53\u9700\u8981\u5177\u5907\u7406 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[8],"tags":[39,95,90,61,80],"class_list":["post-5639","post","type-post","status-publish","format-standard","hentry","category-tech","tag-ai","tag-transformer","tag-multimodal","tag-61","tag-cv"],"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"NullThought","author_link":"https:\/\/nullthought.net\/?author=1"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/nullthought.net\/?cat=8\" rel=\"category\">Tech<\/a>","rttpg_excerpt":"\u8fd1\u5e74\u6765\uff0c\u5177\u8eab\u667a\u80fd\u4f53\uff08Embodied Agents\uff09\u5728\u4eba\u5de5\u667a\u80fd\u7814\u7a76\u4e2d\u53d7\u5230\u4e86\u8d8a\u6765\u8d8a\u591a\u7684\u5173\u6ce8\u3002\u8fd9\u4e9b\u667a\u80fd\u4f53\u9700\u8981\u5177\u5907\u7406&hellip;","_links":{"self":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5639","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=5639"}],"version-history":[{"count":2,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5639\/revisions"}],"predecessor-version":[{"id":5645,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5639\/revisions\/5645"}],"wp:attachment":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5639"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5639"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5639"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}