{"id":25942,"date":"2025-05-21T09:54:30","date_gmt":"2025-05-21T01:54:30","guid":{"rendered":"https:\/\/aif.amtbbs.org\/?p=25942"},"modified":"2025-05-21T09:54:30","modified_gmt":"2025-05-21T01:54:30","slug":"icra-2025%ef%bd%9c%e9%80%9a%e7%94%a8%e5%a4%9a%e6%9c%ba%e5%99%a8%e4%ba%ba%e9%95%bf%e6%97%b6%e4%bb%bb%e5%8a%a1%e8%a7%84%e5%88%92%e6%a1%86%e6%9e%b6%e7%a0%b4%e8%a7%a3%e4%bb%bb%e5%8a%a1%e5%88%86%e9%85%8d","status":"publish","type":"post","link":"https:\/\/aif.amtbbs.org\/index.php\/2025\/05\/21\/25942\/","title":{"rendered":"ICRA 2025\uff5c\u901a\u7528\u591a\u673a\u5668\u4eba\u957f\u65f6\u4efb\u52a1\u89c4\u5212\u6846\u67b6\u7834\u89e3\u4efb\u52a1\u5206\u914d\u96be\u9898\uff0c\u6210\u529f\u7387+105%\u3001\u6548\u7387+36%"},"content":{"rendered":"
\"\"<\/div>\n
<\/div>\n
LaMMA-P \u7684\u63d0\u51fa\u4e3a\u591a\u673a\u5668\u4eba\u957f\u65f6\u4efb\u52a1\u534f\u540c\u89c4\u5212\u63d0\u4f9b\u4e86\u5168\u65b0\u7684\u601d\u8def\uff1a\u5b83\u6210\u529f\u5730\u5c06\u5927\u6a21\u578b\u7684\u667a\u80fd\u63a8\u7406\u4e0e\u7ecf\u5178\u00a0PDDL \u89c4\u5212\u6709\u673a\u7ed3\u5408\uff0c\u5f25\u8865\u4e86\u5404\u81ea\u7684\u77ed\u677f\uff0c\u663e\u8457\u63d0\u5347\u4e86\u590d\u6742\u4efb\u52a1\u7684\u81ea\u52a8\u5316\u7a0b\u5ea6\u548c\u6267\u884c\u53ef\u9760\u6027\u3002<\/div>\n
\n
\n

2025 \u5e74 5 \u6708\uff0c\u7f8e\u56fd\u52a0\u5dde\u5927\u5b66\u6cb3\u6ee8\u5206\u6821 (UC Riverside) \u4e0e\u5bbe\u5915\u6cd5\u5c3c\u4e9a\u5dde\u7acb\u5927\u5b66 (Penn State University) \u8054\u5408\u56e2\u961f\u5728\u673a\u5668\u4eba\u9886\u57df\u9876\u7ea7\u4f1a\u8bae ICRA 2025 \u4e0a\u53d1\u5e03\u6700\u65b0\u7814\u7a76\u6210\u679c\u00a0LaMMA-P\uff08Generalizable Multi-Agent Long-Horizon Task Allocation and Planning with LM-Driven PDDL Planner\uff09\u3002<\/p>\n

LaMMA-P \u9996\u6b21\u5c06\u5927\u578b\u8bed\u8a00\u6a21\u578b\u4e0e PDDL \u89c4\u5212\u5668\u6df1\u5ea6\u878d\u5408\uff0c\u89e3\u51b3\u4e86\u5f02\u6784\u591a\u673a\u5668\u4eba\u7cfb\u7edf\u4e2d\u957f\u65f6\u4efb\u52a1\u7684\u81ea\u52a8\u5206\u89e3\u4e0e\u5206\u914d\u96be\u9898\uff0c\u5927\u5e45\u63d0\u5347\u591a\u673a\u5668\u4eba\u534f\u540c\u89c4\u5212\u7684\u667a\u80fd\u6c34\u5e73\u3002\u8be5\u6280\u672f\u5728\u5168\u65b0\u57fa\u51c6\u6570\u636e\u96c6\u4e0a\u7ecf\u8fc7\u5927\u91cf\u6a21\u62df\u5b9e\u9a8c\u9a8c\u8bc1\uff0c\u76f8\u6bd4\u73b0\u6709\u6700\u5148\u8fdb\u65b9\u6cd5 SMART-LLM\uff0c\u4efb\u52a1\u6210\u529f\u7387\u63d0\u9ad8\u00a0105%\uff0c\u6267\u884c\u6548\u7387\u63d0\u5347\u00a036%\uff0c\u5728\u590d\u6742\u957f\u7a0b\u4efb\u52a1\u89c4\u5212\u4e0a\u53d6\u5f97\u4e86\u7a81\u7834\u6027\u8fdb\u5c55\uff0c\u4e3a\u5f02\u6784\u591a\u673a\u5668\u4eba\u534f\u540c\u5b8c\u6210\u590d\u6742\u4efb\u52a1\u63d0\u4f9b\u4e86\u5168\u65b0\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

    \n
  • \u8bba\u6587\u6807\u9898\uff1a LaMMA-P: Generalizable Multi-Agent Long-Horizon Task Allocation and Planning with LM-Driven PDDL Planner<\/li>\n
  • \u8bba\u6587\u94fe\u63a5\uff1a https:\/\/arxiv.org\/abs\/2409.20560<\/li>\n
  • \u9879\u76ee\u4e3b\u9875\uff1a https:\/\/lamma-p.github.io<\/li>\n
  • \u4ee3\u7801\u5f00\u6e90\uff1a https:\/\/github.com\/tasl-lab\/LaMMA-P<\/li>\n<\/ul>\n

    \u6280\u672f\u4eae\u70b9\uff1a\u8bed\u8a00\u6a21\u578b\u4e0e\u7ecf\u5178\u89c4\u5212\u7b97\u6cd5\u878d\u5408\uff0c\u652f\u6491\u901a\u7528\u5f02\u6784\u591a\u673a\u5668\u4eba\u957f\u65f6\u534f\u540c\u4efb\u52a1<\/p>\n

    \u9762\u5bf9\u590d\u6742\u957f\u65f6\u4efb\u52a1\u548c\u5f02\u6784\u591a\u673a\u5668\u4eba\u7cfb\u7edf\uff0cLaMMA-P \u9996\u521b\u6027\u5730\u5c06\u5927\u8bed\u8a00\u6a21\u578b\u7684\u8bed\u4e49\u7406\u89e3\u80fd\u529b\u4e0e PDDL \u89c4\u5212\u5668\u7684\u4e25\u8c28\u6027\u7ed3\u5408\uff0c\u4e0d\u4ec5\u89e3\u51b3\u4e86\u4f20\u7edf\u65b9\u6cd5\u5728\u4efb\u52a1\u5206\u89e3\u548c\u8d44\u6e90\u5206\u914d\u4e0a\u7684\u74f6\u9888\uff0c\u8fd8\u663e\u8457\u63d0\u5347\u4e86\u4efb\u52a1\u6210\u529f\u7387\u4e0e\u6267\u884c\u6548\u7387\u3002\u901a\u8fc7\u5168\u9762\u5f00\u6e90\u7684\u4ee3\u7801\u4e0e MAT-THOR \u57fa\u51c6\u6570\u636e\u96c6\uff0cLaMMA-P \u4e3a\u591a\u673a\u5668\u4eba\u534f\u540c\u89c4\u5212\u5f00\u8f9f\u4e86\u65b0\u7684\u6280\u672f\u8def\u5f84\uff1a<\/p>\n

      \n
    1. \u5927\u6a21\u578b\u9a71\u52a8\u7684 PDDL \u89c4\u5212\u6846\u67b6\uff1a\u63d0\u51fa\u5c06\u5927\u578b\u8bed\u8a00\u6a21\u578b (Large Language Model, LLM) \u7684\u5f3a\u5927\u4efb\u52a1\u7406\u89e3\u548c\u63a8\u7406\u80fd\u529b\u4e0e\u7ecf\u5178\u89c4\u5212\u9886\u57df\u5b9a\u4e49\u8bed\u8a00 (Planning Domain Definition Language, PDDL) \u89c4\u5212\u5668\u76f8\u7ed3\u5408\u7684\u65b0\u6846\u67b6\u3002\u5229\u7528 LLM \u5c06\u4eba\u7c7b\u6307\u4ee4\u8f6c\u5316\u4e3a\u9ad8\u5c42\u4efb\u52a1\u63cf\u8ff0\uff0c\u518d\u7531 PDDL \u89c4\u5212\u5668\u8fdb\u884c\u4e25\u8c28\u7684\u5b50\u4efb\u52a1\u641c\u7d22\u548c\u89c4\u5212\uff0c\u5b9e\u73b0\u5bf9\u957f\u65f6\u590d\u6742\u4efb\u52a1\u7684\u81ea\u52a8\u5206\u89e3\u4e0e\u5206\u914d\u3002\u8be5\u65b9\u6cd5\u878d\u5408\u4e86\u5b66\u4e60\u5f0f\u63a8\u7406\u4e0e\u542f\u53d1\u5f0f\u641c\u7d22\u7684\u4f18\u52bf\uff0c\u65e2\u80fd\u7406\u89e3\u590d\u6742\u81ea\u7136\u8bed\u8a00\u6307\u4ee4\uff0c\u53c8\u80fd\u63d0\u4f9b\u53ef\u9760\u9ad8\u6548\u7684\u89c4\u5212\u65b9\u6848\u3002<\/li>\n
    2. \u6a21\u5757\u5316\u8bbe\u8ba1\u4e0e\u5f3a\u6cdb\u5316\u80fd\u529b\uff1aLaMMA-P \u91c7\u7528\u6a21\u5757\u5316\u67b6\u6784\uff0c\u5c06\u5927\u578b\u8bed\u8a00\u6a21\u578b\u3001PDDL \u89c4\u5212\u7cfb\u7edf\u4e0e\u4eff\u771f\u73af\u5883\u65e0\u7f1d\u96c6\u6210\u3002\u901a\u8fc7\u6a21\u5757\u95f4\u6e05\u6670\u7684\u63a5\u53e3\uff0c\u7cfb\u7edf\u53ef\u6839\u636e\u4e0d\u540c\u673a\u5668\u4eba\u6280\u80fd\u7075\u6d3b\u5206\u89e3\u4efb\u52a1\u5e76\u9ad8\u6548\u5206\u914d\u5b50\u4efb\u52a1\u3002\u8be5\u6846\u67b6\u652f\u6301\u4efb\u610f\u6570\u91cf\u7684\u673a\u5668\u4eba\u53c2\u4e0e\uff0c\u540c\u4e00\u5957\u7b97\u6cd5\u65e0\u9700\u4fee\u6539\u5373\u53ef\u6269\u5c55\u5230\u66f4\u591a\u673a\u5668\u4eba\u534f\u4f5c\u573a\u666f\uff0c\u4f53\u73b0\u51fa\u4f18\u5f02\u7684\u6cdb\u5316\u80fd\u529b\u3002\u5728\u4e0d\u540c\u5bb6\u5ead\u4efb\u52a1\u3001\u4e0d\u540c\u673a\u5668\u4eba\u7ec4\u5408\u4e0b\uff0cLaMMA-P \u90fd\u80fd\u4ea7\u51fa\u6709\u6548\u7684\u534f\u540c\u8ba1\u5212\uff0c\u8868\u73b0\u51fa\u5bf9\u4efb\u52a1\u79cd\u7c7b\u548c\u56e2\u961f\u89c4\u6a21\u7684\u5f3a\u9002\u5e94\u6027\u3002<\/li>\n
    3. \u65b0\u57fa\u51c6\u6570\u636e\u96c6\u4e0e\u6027\u80fd\u8d85\u8d8a\uff1a\u6784\u5efa\u4e86\u5168\u65b0\u7684\u591a\u667a\u80fd\u4f53\u957f\u65f6\u4efb\u52a1\u6a21\u62df\u57fa\u51c6\u00a0MAT-THOR\uff0c\u57fa\u4e8e AI2-THOR \u4eff\u771f\u73af\u5883\u8bbe\u8ba1\u5bb6\u5ead\u573a\u666f\u4e0b\u7684\u590d\u6742\u534f\u540c\u4efb\u52a1\uff08\u5305\u542b\u7b80\u5355\u590d\u5408\u4efb\u52a1\u548c\u9ad8\u590d\u6742\u4efb\u52a1\u4e24\u79cd\u7ea7\u522b\uff09\u3002\u8be5\u6570\u636e\u96c6\u4e3a\u591a\u673a\u5668\u4eba\u957f\u7a0b\u4efb\u52a1\u89c4\u5212\u63d0\u4f9b\u4e86\u6807\u51c6\u8bc4\u6d4b\u5e73\u53f0\u3002\u5b9e\u9a8c\u7ed3\u679c\u663e\u793a\uff0cLaMMA-P \u5728 MAT-THOR \u57fa\u51c6\u4e0a\u53d6\u5f97\u4e86\u5f53\u524d\u6700\u4f18\u6210\u7ee9\uff1a\u76f8\u8f83\u4e8e\u6700\u65b0\u7684\u5927\u6a21\u578b\u591a\u673a\u5668\u4eba\u89c4\u5212\u65b9\u6848 SMART-LLM\uff0cLaMMA-P \u7684\u4efb\u52a1\u5b8c\u6210\u6210\u529f\u7387\u63d0\u9ad8\u4e86 105%\uff0c\u6548\u7387\u63d0\u5347\u4e86 36%\uff0c\u5728\u957f\u65f6\u4efb\u52a1\u7684\u6210\u529f\u6267\u884c\u7387\u548c\u6267\u884c\u901f\u5ea6\u4e24\u65b9\u9762\u5747\u5927\u5e45\u9886\u5148\u73b0\u6709\u65b9\u6cd5\u3002<\/li>\n<\/ol>\n

      \u4e00\uff0e\u7814\u7a76\u80cc\u666f<\/h4>\n

      \u968f\u7740\u591a\u673a\u5668\u4eba\u7cfb\u7edf\u5728\u641c\u6551\u3001\u4ed3\u50a8\u3001\u5bb6\u5ead\u670d\u52a1\u7b49\u573a\u666f\u4e2d\u65e5\u76ca\u666e\u53ca\uff0c\u8ba9\u591a\u53f0\u5f02\u6784\u673a\u5668\u4eba\u534f\u540c\u6267\u884c\u590d\u6742\u4efb\u52a1\u6210\u4e3a\u673a\u5668\u4eba\u9886\u57df\u7684\u91cd\u8981\u8bfe\u9898\u3002\u5728\u73b0\u5b9e\u5e94\u7528\u4e2d\uff0c\u957f\u65f6\u4efb\u52a1\u901a\u5e38\u9700\u8981\u88ab\u5206\u89e3\u4e3a\u4e00\u7cfb\u5217\u5173\u8054\u7684\u5b50\u4efb\u52a1\uff0c\u5e76\u5408\u7406\u5206\u914d\u7ed9\u5177\u5907\u4e0d\u540c\u80fd\u529b\u7684\u673a\u5668\u4eba\u4e2a\u4f53\u534f\u540c\u5b8c\u6210\u3002\u4f8b\u5982\u5bb6\u5ead\u52a9\u7406\u673a\u5668\u4eba\u9700\u8981\u5408\u4f5c\u5b8c\u6210 \u201c\u6574\u7406\u623f\u95f4\u201d \u8fd9\u6837\u7684\u590d\u6742\u6307\u4ee4\uff0c\u5176\u4e2d\u5305\u542b\u642c\u8fd0\u7269\u54c1\u3001\u6e05\u6d01\u3001\u6574\u7406\u7b49\u591a\u4e2a\u6b65\u9aa4\u3002\u7136\u800c\uff0c\u957f\u65f6\u4efb\u52a1\u7684\u81ea\u52a8\u5206\u89e3\u4e0e\u591a\u673a\u5668\u4eba\u5206\u5de5\u6781\u5177\u6311\u6218\uff1a\u5982\u4f55\u8ba9\u673a\u5668\u4eba\u961f\u4f0d\u7406\u89e3\u4eba\u7c7b\u7684\u9ad8\u5c42\u610f\u56fe\uff0c\u5c06\u5176\u7ec6\u5316\u4e3a\u53ef\u6267\u884c\u7684\u5b50\u4efb\u52a1\u6e05\u5355\uff0c\u5e76\u6839\u636e\u6bcf\u53f0\u673a\u5668\u4eba\u7684\u80fd\u529b\u8fdb\u884c\u6700\u4f73\u5206\u914d\uff0c\u662f\u5f53\u524d\u667a\u80fd\u4f53\u534f\u4f5c\u9762\u4e34\u7684\u96be\u70b9\u3002<\/p>\n

      \u4f20\u7edf\u7684\u591a\u673a\u5668\u4eba\u4efb\u52a1\u89c4\u5212\u65b9\u6cd5\u5f80\u5f80\u4f9d\u8d56\u9884\u5148\u5b9a\u4e49\u7684\u89c4\u5219\u6216\u542f\u53d1\u5f0f\u7b97\u6cd5\uff0c\u5728\u4efb\u52a1\u89c4\u6a21\u548c\u590d\u6742\u5ea6\u8f83\u5c0f\u65f6\u5c1a\u80fd\u594f\u6548\uff0c\u4f46\u9762\u5bf9\u5177\u6709\u957f\u65f6\u95f4\u8de8\u5ea6\u3001\u590d\u6742\u4f9d\u8d56\u5173\u7cfb\u7684\u4efb\u52a1\u65f6\u5f80\u5f80\u529b\u4e0d\u4ece\u5fc3\u3002\u8fd9\u7c7b\u65b9\u6cd5\u7f3a\u4e4f\u5bf9\u81ea\u7136\u8bed\u8a00\u6307\u4ee4\u8bed\u4e49\u7684\u6df1\u5165\u7406\u89e3\uff0c\u96be\u4ee5\u53ca\u65f6\u8c03\u6574\u89c4\u5212\u6765\u5e94\u5bf9\u52a8\u6001\u73af\u5883\u548c\u4efb\u52a1\u53d8\u5316\u3002\u6b64\u5916\uff0c\u57fa\u4e8e\u5927\u578b\u8bed\u8a00\u6a21\u578b\u7684\u673a\u5668\u4eba\u4efb\u52a1\u89c4\u5212\u867d\u7136\u5728\u5355\u673a\u5668\u4eba\u7684\u7b80\u5355\u4efb\u52a1\u4e0a\u53d6\u5f97\u4e86\u4e00\u5b9a\u8fdb\u5c55\uff0c\u4f46\u5728\u591a\u673a\u5668\u4eba\u534f\u540c\u7684\u957f\u7a0b\u4efb\u52a1\u4e0a\u4f9d\u7136\u5b58\u5728\u6210\u529f\u7387\u4f4e\u3001\u6548\u7387\u5dee\u3001\u96be\u4ee5\u6cdb\u5316\u7b49\u95ee\u9898\u3002\u4e3a\u4e86\u89e3\u51b3\u4e0a\u8ff0\u74f6\u9888\uff0c\u7814\u7a76\u4eba\u5458\u5f00\u59cb\u63a2\u7d22\u5c06\u65b0\u5174\u7684\u5927\u8bed\u8a00\u6a21\u578b\u4e0e\u7ecf\u5178\u4eba\u5de5\u667a\u80fd\u89c4\u5212\u6280\u672f\u76f8\u7ed3\u5408\uff0c\u8bd5\u56fe\u517c\u987e\u4e24\u8005\u4f18\u52bf\u6765\u63d0\u5347\u591a\u673a\u5668\u4eba\u7cfb\u7edf\u7684\u81ea\u4e3b\u534f\u540c\u80fd\u529b\u3002LaMMA-P\u00a0\u6b63\u662f\u5728\u8fd9\u6837\u7684\u80cc\u666f\u4e0b\u5e94\u8fd0\u800c\u751f\u3002<\/p>\n

      \u4e8c\uff0e\u7814\u7a76\u65b9\u6cd5<\/h4>\n

      \u5927\u578b\u8bed\u8a00\u6a21\u578b\u548c\u542f\u53d1\u5f0f\u641c\u7d22\u89c4\u5212\u5668\u5171\u540c\u9a71\u52a8\u7684\u591a\u673a\u5668\u4eba\u89c4\u5212\u6846\u67b6<\/h4>\n

      \"\u56fe\u7247\"<\/p>\n

      \u9762\u5bf9\u5f02\u6784\u591a\u673a\u5668\u4eba\u957f\u65f6\u4efb\u52a1\u89c4\u5212\u96be\u9898\uff0cLaMMA-P \u63d0\u51fa\u4e86\u878d\u5408\u5927\u578b\u8bed\u8a00\u6a21\u578b\u4e0e PDDL \u89c4\u5212\u7684\u521b\u65b0\u6846\u67b6\u3002\u9996\u5148\uff0c\u7cfb\u7edf\u5229\u7528\u5927\u8bed\u8a00\u6a21\u578b\u5f3a\u5927\u7684\u63a8\u7406\u4e0e\u6587\u672c\u89e3\u6790\u80fd\u529b\u6765\u7406\u89e3\u590d\u6742\u7684\u81ea\u7136\u8bed\u8a00\u6307\u4ee4\uff0c\u4ece\u4e2d\u62bd\u53d6\u9ad8\u5c42\u4efb\u52a1\u76ee\u6807\u548c\u7ea6\u675f\uff0c\u5e76\u751f\u6210\u521d\u6b65\u7684\u4efb\u52a1\u5206\u89e3\u65b9\u6848\u3002\u63a5\u7740\uff0cLaMMA-P \u5c06\u4efb\u52a1\u5206\u89e3\u7ed3\u679c\u8f6c\u6362\u4e3a\u5f62\u5f0f\u5316\u7684 PDDL \u8868\u8ff0\uff0c\u7531\u7ecf\u5178\u89c4\u5212\u5668\u6267\u884c\u5168\u5c40\u641c\u7d22\u4e0e\u6700\u4f18\u5b50\u4efb\u52a1\u89c4\u5212\uff0c\u627e\u51fa\u6ee1\u8db3\u7ea6\u675f\u7684\u6700\u4f18\u6267\u884c\u5e8f\u5217\u548c\u5206\u914d\u65b9\u6848\u3002\u5728\u8fd9\u4e2a\u8fc7\u7a0b\u4e2d\uff0c\u5927\u8bed\u8a00\u6a21\u578b\u8d1f\u8d23\u8bed\u4e49\u7406\u89e3\u548c\u9ad8\u5c42\u51b3\u7b56\uff0cPDDL \u89c4\u5212\u5668\u4fdd\u8bc1\u4e86\u89c4\u5212\u7ed3\u679c\u7684\u4e25\u8c28\u6027\u548c\u5b8c\u5907\u6027\u3002\u901a\u8fc7\u4eba\u673a\u6307\u4ee4\u7406\u89e3\u548c\u5f62\u5f0f\u89c4\u5212\u6c42\u89e3\u7684\u7ed3\u5408\uff0cLaMMA-P \u5b9e\u73b0\u4e86\u5bf9\u957f\u65f6\u4efb\u52a1\u65e2\u80fd\u6e05\u695a\u7406\u89e3\u53c8\u80fd\u7cbe\u786e\u8ba1\u7b97\u3002<\/p>\n

      \u6574\u4e2a\u6846\u67b6\u91c7\u7528\u6a21\u5757\u5316\u8bbe\u8ba1\uff0c\u5305\u62ec\u6307\u4ee4\u89e3\u6790\u6a21\u5757\u3001\u4efb\u52a1\u89c4\u5212\u6a21\u5757\u3001\u6267\u884c\u4eff\u771f\u6a21\u5757\u7b49\u3002\u6307\u4ee4\u89e3\u6790\u6a21\u5757\u7531\u9884\u8bad\u7ec3\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b\u5b9e\u73b0\uff0c\u8f93\u51fa\u4efb\u52a1\u7ed3\u6784\u56fe\u548c\u521d\u6b65\u5206\u914d\u5efa\u8bae\uff1b\u4efb\u52a1\u89c4\u5212\u6a21\u5757\u57fa\u4e8e PDDL\uff0c\u901a\u8fc7\u8fed\u4ee3\u6df1\u5316\u641c\u7d22\u5f97\u5230\u4efb\u52a1\u5206\u914d\u7684\u5177\u4f53\u884c\u52a8\u5e8f\u5217\uff1b\u6267\u884c\u6a21\u5757\u5728\u4eff\u771f\u73af\u5883\u4e2d\u9a8c\u8bc1\u5e76\u53cd\u9988\u6267\u884c\u7ed3\u679c\u7528\u4e8e\u53ef\u80fd\u7684\u8c03\u6574\u3002\u5404\u6a21\u5757\u89e3\u8026\u7684\u8bbe\u8ba1\u4f7f\u5f97\u589e\u52a0\u673a\u5668\u4eba\u6570\u91cf\u6216\u66f4\u6539\u4efb\u52a1\u7c7b\u578b\u65f6\u65e0\u9700\u66f4\u6539\u6846\u67b6\u4e3b\u4f53\u903b\u8f91\uff0c\u53ea\u9700\u8c03\u6574\u76f8\u5e94\u7684\u4efb\u52a1\u63cf\u8ff0\u548c\u673a\u5668\u4eba\u80fd\u529b\u914d\u7f6e\uff0c\u4f53\u73b0\u51fa\u9ad8\u5ea6\u7684\u7075\u6d3b\u6027\u548c\u901a\u7528\u6027\u3002<\/p>\n

      \u4e09\uff0e\u5b9e\u9a8c\u7ed3\u679c<\/h4>\n

      \u6027\u80fd\u5168\u9762\u8d85\u8d8a\u73b0\u6709\u65b9\u6848<\/h4>\n

      \u7814\u7a76\u56e2\u961f\u6784\u5efa\u4e86\u00a0MAT-THOR\u00a0\u6570\u636e\u96c6\u6765\u8bc4\u4f30 LaMMA-P \u7684\u6709\u6548\u6027\u3002MAT-THOR \u57fa\u51c6\u6db5\u76d6\u4e86\u5728\u5bb6\u5ead\u573a\u666f\u4e0b\u591a\u673a\u5668\u4eba\u534f\u4f5c\u5b8c\u6210\u957f\u65f6\u4efb\u52a1\u7684\u591a\u79cd\u60c5\u51b5\uff0c\u5305\u62ec\u7b80\u5355\u7684\u987a\u5e8f\u4efb\u52a1\u548c\u590d\u6742\u7684\u5e76\u884c\u4efb\u52a1\uff0c\u6a21\u62df\u771f\u5b9e\u5bb6\u5ead\u73af\u5883\u4e2d\u7684\u591a\u6837\u6311\u6218\u3002\u5b9e\u9a8c\u5728\u4e0d\u540c\u6570\u91cf\u673a\u5668\u4eba\uff08\u5305\u62ec 2 \u53f0\u53ca\u66f4\u591a\uff09\u548c\u4e0d\u540c\u4efb\u52a1\u590d\u6742\u5ea6\u7ec4\u5408\u4e0b\u8fdb\u884c\uff0c\u5e76\u5c06 LaMMA-P \u4e0e\u5f53\u524d\u6700\u65b0\u7684\u591a\u673a\u5668\u4eba\u4efb\u52a1\u89c4\u5212\u65b9\u6cd5\u00a0SMART-LLM\u00a0\u8fdb\u884c\u4e86\u5bf9\u6bd4\u3002\u5b9e\u9a8c\u4ece\u6027\u80fd\u5bf9\u6bd4\u3001\u6a21\u5757\u6d88\u878d\u548c\u53ef\u89c6\u5316\u5206\u6790\u4e09\u4e2a\u65b9\u9762\u7cfb\u7edf\u5c55\u793a\u4e86 LaMMA-P \u7684\u4f18\u52bf\u3002<\/p>\n

      1. \u6027\u80fd\u5bf9\u6bd4<\/strong><\/p>\n

      LaMMA-P \u5728\u4efb\u52a1\u6210\u529f\u7387\u3001\u6548\u7387\u4ee5\u53ca\u673a\u5668\u4eba\u5229\u7528\u7387\u7b49\u591a\u4e2a\u6838\u5fc3\u6307\u6807\u4e0a\u5747\u5927\u5e45\u9886\u5148\u5f53\u524d\u6700\u4f18\u65b9\u6cd5\u00a0SMART-LLM (GPT-4o)\uff1a<\/p>\n

      \"\u56fe\u7247\"<\/p>\n

      \u5173\u952e\u53d1\u73b0\uff1a<\/p>\n

        \n
      • \u957f\u65f6\u4efb\u52a1\u6210\u529f\u7387\u63d0\u5347\u663e\u8457\uff1a\u5728 Compound \u4efb\u52a1\u4e2d\uff0c\u6210\u529f\u7387\u63d0\u5347\u81f3\u00a093%\uff08\u9886\u5148 23%\uff09\u3002Complex \u4efb\u52a1\u4e2d\uff0c\u6210\u529f\u7387\u63d0\u5347\u81f3\u00a077%\uff08\u9886\u5148 57%\uff09\u3002\u5728 Vague Command \u4efb\u52a1\u4e2d\uff0c\u6210\u529f\u7387\u8fbe\u5230\u00a045%\uff0c\u800c\u73b0\u6709\u65b9\u6cd5\u5b8c\u5168\u5931\u8d25\u3002<\/li>\n
      • \u6548\u7387\u4e0e\u673a\u5668\u4eba\u5229\u7528\u7387\u5168\u9762\u63d0\u5347\uff1a\u6267\u884c\u6548\u7387\uff08Eff\uff09\u5728\u4e09\u7c7b\u4efb\u52a1\u4e2d\u5206\u522b\u63d0\u5347\u00a026% ~ 48%\u3002\u673a\u5668\u4eba\u5229\u7528\u7387\uff08RU\uff09\u5728 Complex \u4efb\u52a1\u4e2d\u63d0\u5347\u00a022%\uff0c\u663e\u8457\u4f18\u5316\u591a\u673a\u5668\u4eba\u534f\u4f5c\u8c03\u5ea6\u3002\u6267\u884c\u52a8\u4f5c\u53ef\u884c\u6027\uff08Exe\uff09\u8fbe\u5230\u00a0100%\uff0c\u4fdd\u8bc1\u4efb\u52a1\u5206\u914d\u540e\u6bcf\u4e00\u6b65\u90fd\u80fd\u987a\u5229\u5b8c\u6210\u3002<\/li>\n
      • \u590d\u6742\u4efb\u52a1\u4e2d\u4f18\u52bf\u66f4\u52a0\u7a81\u51fa\uff1a\u968f\u7740\u4efb\u52a1\u590d\u6742\u5ea6\u589e\u52a0\uff0cLaMMA-P \u4f18\u52bf\u6108\u53d1\u660e\u663e\uff0c\u8868\u73b0\u51fa\u5353\u8d8a\u7684\u6cdb\u5316\u548c\u63a8\u7406\u80fd\u529b\u3002Vague Command \u573a\u666f\u4e2d\u552f\u4e00\u80fd\u591f\u6210\u529f\u5b8c\u6210\u4efb\u52a1\u7684\u65b9\u6cd5\uff0c\u5c55\u73b0\u5f3a\u5927\u7684\u81ea\u7136\u8bed\u8a00\u7406\u89e3\u548c\u6a21\u7cca\u6307\u4ee4\u89e3\u6790\u80fd\u529b\u3002<\/li>\n<\/ul>\n

        2. \u6a21\u5757\u6d88\u878d\u5206\u6790<\/strong><\/p>\n

        \u901a\u8fc7\u6d88\u878d\u4e0d\u540c\u6a21\u5757\uff0c\u9a8c\u8bc1\u4e86 LaMMA-P \u5404\u6a21\u5757\u5bf9\u6574\u4f53\u6027\u80fd\u7684\u5173\u952e\u8d21\u732e\uff1a<\/p>\n

        \"\u56fe\u7247\"<\/p>\n

        \u5173\u952e\u53d1\u73b0\uff1a<\/p>\n

          \n
        • Precondition Identifier (P) \u6a21\u5757\u663e\u8457\u63d0\u5347\u89c4\u5212\u6210\u529f\u7387\uff1a\u79fb\u9664 P \u6a21\u5757\u540e\uff0cCompound \u4efb\u52a1\u6210\u529f\u7387\u4e0b\u964d\u00a014%\uff0cComplex \u4efb\u52a1\u4e0b\u964d\u00a09%\u3002P \u6a21\u5757\u5e2e\u52a9\u6a21\u578b\u66f4\u597d\u89e3\u6790\u5148\u51b3\u6761\u4ef6\uff0c\u63d0\u9ad8\u5b50\u4efb\u52a1\u5408\u7406\u6027\u548c\u8ba1\u5212\u6210\u529f\u7387\u3002<\/li>\n
        • PDDL Validator (V) \u548c Problem Generator (G) \u63d0\u9ad8\u6267\u884c\u53ef\u9760\u6027\uff1a\u53bb\u9664 V \u548c G \u540e\uff0cComplex \u4efb\u52a1\u6267\u884c\u7387\u4e0b\u964d\u81f3\u00a087%\uff08\u76f8\u6bd4\u5b8c\u6574\u7cfb\u7edf\u7684 100%\uff09\u3002V \u6a21\u5757\u786e\u4fdd\u751f\u6210\u7684 PDDL \u95ee\u9898\u683c\u5f0f\u6b63\u786e\uff0c\u6709\u6548\u907f\u514d\u56e0\u4e0d\u5408\u89c4\u8ba1\u5212\u5bfc\u81f4\u7684\u5931\u8d25\u3002<\/li>\n
        • \u5b8c\u6574\u6a21\u5757\u7ec4\u5408\u5e26\u6765\u6700\u4f18\u6027\u80fd\uff1a\u5305\u542b\u5168\u90e8\u6a21\u5757\u7684 LaMMA-P \u5b9e\u73b0\u6700\u4f73\u6548\u679c\uff0cCompound \u4efb\u52a1\u6210\u529f\u7387\u9ad8\u8fbe\u00a093%\uff0cComplex \u4efb\u52a1\u6210\u529f\u7387\u8fbe\u00a077%\u3002\u6574\u4f53\u6548\u7387\uff08Eff\uff09\u63d0\u5347\u6700\u660e\u663e\uff0c\u8868\u660e\u5b8c\u6574\u7cfb\u7edf\u80fd\u591f\u4f18\u5316\u4efb\u52a1\u5206\u89e3\u3001\u8d44\u6e90\u8c03\u5ea6\u4e0e\u8ba1\u5212\u6267\u884c\u7684\u5168\u6d41\u7a0b\u3002<\/li>\n<\/ul>\n

          3. \u53ef\u89c6\u5316\u5206\u6790<\/strong><\/p>\n

          \u901a\u8fc7 AI2-THOR \u4eff\u771f\u73af\u5883\u53ef\u89c6\u5316\u5c55\u793a\uff0c\u76f4\u89c2\u4f53\u73b0 LaMMA-P \u4f18\u5316\u7684\u4efb\u52a1\u89c4\u5212\u6548\u679c\uff1a<\/p>\n

          \"\u56fe\u7247\"<\/p>\n

          \u5173\u952e\u53d1\u73b0\uff1a<\/p>\n

            \n
          • \u5e76\u884c\u4e0e\u987a\u5e8f\u4efb\u52a1\u5408\u7406\u8c03\u5ea6\uff1a\u5728 Compound \u4efb\u52a1\u4e2d\uff0c\u591a\u4e2a\u673a\u5668\u4eba\u57fa\u4e8e\u81ea\u8eab\u80fd\u529b\u5e76\u884c\u6267\u884c\u4e0d\u540c\u5b50\u4efb\u52a1\uff0c\u63d0\u9ad8\u6574\u4f53\u4efb\u52a1\u5b8c\u6210\u901f\u5ea6\u3002\u901a\u8fc7\u667a\u80fd\u987a\u5e8f\u5b89\u6392\uff0c\u907f\u514d\u56e0\u573a\u666f\u62e5\u5835\u6216\u4efb\u52a1\u51b2\u7a81\u5bfc\u81f4\u7684\u6267\u884c\u5ef6\u8fdf\u3002<\/li>\n
          • \u590d\u6742\u4f9d\u8d56\u5173\u7cfb\u4e0b\u7684\u7cbe\u51c6\u534f\u4f5c\uff1a\u5728 Complex \u4efb\u52a1\u4e2d\uff0c\u673a\u5668\u4eba\u80fd\u591f\u5408\u7406\u7b49\u5f85\u5fc5\u8981\u6761\u4ef6\u6ee1\u8db3\uff08\u5982\u7b49\u5f85\u62bd\u5c49\u88ab\u6253\u5f00\uff09\uff0c\u4f53\u73b0\u5bf9\u4efb\u52a1\u5148\u540e\u7ea6\u675f\u7684\u7cbe\u51c6\u628a\u63a7\u3002\u6709\u6548\u51cf\u5c11\u65e0\u6548\u7b49\u5f85\u548c\u91cd\u590d\u79fb\u52a8\uff0c\u63d0\u9ad8\u673a\u5668\u4eba\u6574\u4f53\u5229\u7528\u7387\u3002<\/li>\n<\/ul>\n
            \u56db\uff0e\u603b\u7ed3\u4e0e\u5c55\u671b<\/h5>\n
            \u591a\u673a\u5668\u4eba\u534f\u540c\u89c4\u5212\u65b0\u8303\u5f0f<\/h5>\n

            LaMMA-P \u7684\u63d0\u51fa\u4e3a\u591a\u673a\u5668\u4eba\u957f\u65f6\u4efb\u52a1\u534f\u540c\u89c4\u5212\u63d0\u4f9b\u4e86\u5168\u65b0\u7684\u601d\u8def\uff1a\u5b83\u6210\u529f\u5730\u5c06\u5927\u6a21\u578b\u7684\u667a\u80fd\u63a8\u7406\u4e0e\u7ecf\u5178\u00a0PDDL \u89c4\u5212\u6709\u673a\u7ed3\u5408\uff0c\u5f25\u8865\u4e86\u5404\u81ea\u7684\u77ed\u677f\uff0c\u663e\u8457\u63d0\u5347\u4e86\u590d\u6742\u4efb\u52a1\u7684\u81ea\u52a8\u5316\u7a0b\u5ea6\u548c\u6267\u884c\u53ef\u9760\u6027\u3002\u8fd9\u4e00\u6210\u679c\u6807\u5fd7\u7740\u5f02\u6784\u591a\u673a\u5668\u4eba\u81ea\u4e3b\u534f\u4f5c\u8fc8\u51fa\u4e86\u5173\u952e\u4e00\u6b65\u3002\u5728\u672a\u6765\u5de5\u4f5c\u4e2d\uff0c\u7814\u7a76\u56e2\u961f\u8ba1\u5212\u8fdb\u4e00\u6b65\u63a2\u7d22\u7aef\u5230\u7aef\u7684\u4f18\u5316\uff08\u4f8b\u5982\u4f7f\u8bed\u8a00\u6a21\u578b\u4e0e\u89c4\u5212\u5668\u8054\u7ed3\u5f62\u6210\u53ef\u5fae\u5206\u95ed\u73af\uff09\uff0c\u4ee5\u53ca\u5f15\u5165\u591a\u6a21\u6001\u611f\u77e5\u4fe1\u606f\uff08\u5982\u89c6\u89c9\u3001\u8bed\u97f3\uff09\u6765\u4e30\u5bcc\u6307\u4ee4\u7406\u89e3\uff0c\u4e0d\u65ad\u63d0\u5347\u7cfb\u7edf\u5728\u771f\u5b9e\u73af\u5883\u4e2d\u7684\u9002\u5e94\u6027\u548c\u9c81\u68d2\u6027\u3002\u968f\u7740\u4ee3\u7801\u3001\u6570\u636e\u96c6\u548c\u4ea4\u4e92\u5f0f\u6f14\u793a\u7684\u5f00\u6e90\u53d1\u5e03\uff0c\u66f4\u591a\u7814\u7a76\u8005\u548c\u5f00\u53d1\u8005\u53ef\u4ee5\u57fa\u4e8e LaMMA-P \u6846\u67b6\u8fdb\u884c\u62d3\u5c55\u521b\u65b0\uff0c\u52a0\u901f\u591a\u673a\u5668\u4eba\u534f\u540c\u9886\u57df\u7684\u6280\u672f\u6f14\u8fdb\u3002\u53ef\u4ee5\u9884\u89c1\uff0cLaMMA-P \u6240\u4ee3\u8868\u7684 \u201c\u5927\u6a21\u578b\u9a71\u52a8\u89c4\u5212\u201d \u65b0\u8303\u5f0f\u6709\u671b\u5728\u672a\u6765\u8d70\u51fa\u5b9e\u9a8c\u5ba4\uff0c\u5e94\u7528\u5230\u667a\u80fd\u5236\u9020\u3001\u5bb6\u5ead\u670d\u52a1\u7b49\u4f17\u591a\u573a\u666f\uff0c\u63a8\u52a8\u591a\u673a\u5668\u4eba\u7cfb\u7edf\u671d\u7740\u66f4\u52a0\u9ad8\u6548\u3001\u667a\u80fd\u7684\u65b9\u5411\u53d1\u5c55\u3002<\/p>\n

            \u6587\u7ae0\u6765\u81ea\uff1a51CTO<\/p>\n<\/div>\n<\/div>\n

            <\/div>\n

            <\/g><\/svg><\/i> \"Loading\"<\/p>\n

            <\/div>\n","protected":false},"excerpt":{"rendered":"

            LaMMA-P \u7684\u63d0\u51fa\u4e3a\u591a\u673a\u5668\u4eba\u957f\u65f6\u4efb\u52a1\u534f\u540c\u89c4\u5212\u63d0\u4f9b\u4e86\u5168\u65b0\u7684\u601d\u8def\uff1a\u5b83\u6210\u529f\u5730\u5c06\u5927\u6a21\u578b\u7684\u667a\u80fd\u63a8\u7406\u4e0e\u7ecf\u5178\u00a0PDDL […]<\/p>\n

            <\/div>\n

            <\/g><\/svg><\/i> \"Loading\"<\/p>\n

            <\/div>\n","protected":false},"author":56,"featured_media":25944,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[23,20,80,7],"tags":[124,112,820],"class_list":["post-25942","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-23","category-20","category-80","category-7","tag-ai","tag-112","tag-820"],"_links":{"self":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/25942","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=25942"}],"version-history":[{"count":1,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/25942\/revisions"}],"predecessor-version":[{"id":25945,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/25942\/revisions\/25945"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media\/25944"}],"wp:attachment":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media?parent=25942"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/categories?post=25942"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/tags?post=25942"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}