{"id":22629,"date":"2024-10-16T11:27:55","date_gmt":"2024-10-16T03:27:55","guid":{"rendered":"https:\/\/aif.amtbbs.org\/?p=22629"},"modified":"2024-10-16T11:27:55","modified_gmt":"2024-10-16T03:27:55","slug":"%e5%ae%8c%e5%85%a8%e5%9c%b0%e7%90%86%e8%a7%a3%e8%bf%99%e4%b8%aa%e4%b8%96%e7%95%8c%e6%98%af%e4%b8%96%e7%95%8c%e6%a8%a1%e5%9e%8b%e8%a6%81%e5%b9%b2%e7%9a%84%e4%ba%8b%ef%bc%812024%e8%87%aa%e5%8a%a8","status":"publish","type":"post","link":"https:\/\/aif.amtbbs.org\/index.php\/2024\/10\/16\/22629\/","title":{"rendered":"\u5b8c\u5168\u5730\u7406\u89e3\u8fd9\u4e2a\u4e16\u754c\u662f\u4e16\u754c\u6a21\u578b\u8981\u5e72\u7684\u4e8b\uff012024\u81ea\u52a8\u9a7e\u9a76\u4e16\u754c\u6a21\u578b\u5927\u89c2~"},"content":{"rendered":"
\"\"<\/div>\n
<\/div>\n
\u5f53\u4e0b\u81ea\u52a8\u9a7e\u9a76\u65b9\u5411\u7684\u4e16\u754c\u6a21\u578b\u53ef\u4ee5\u5206\u6210\u4e24\u5927\u7c7b\uff1a\u751f\u6210\u5f0f\u548c\u7aef\u5230\u7aef\u3002<\/div>\n
\n
\n

\u672c\u6587\u7ecf\u81ea\u52a8\u9a7e\u9a76\u4e4b\u5fc3\u516c\u4f17\u53f7\u6388\u6743\u8f6c\u8f7d\uff0c\u8f6c\u8f7d\u8bf7\u8054\u7cfb\u51fa\u5904\u3002<\/p>\n

\u4f55\u8c13\u4e16\u754c\u6a21\u578b\uff1f<\/h2>\n

\u201c\u6574\u4f53\u4e0a\u6765\u8bf4\uff0c\u5b8c\u5168\u5730\u7406\u89e3\u8fd9\u4e2a\u4e16\u754c\u662f\u4e16\u754c\u6a21\u578b\u8981\u5e72\u7684\u4e8b\u3002\u201d\u2014\u2014\u4efb\u5c11\u537f\u5728\u63a5\u53d7\u91c7\u8bbf\u4e2d\u8bf4\u5230\u3002<\/p>\n

\u90a3\u4e48\u4f55\u8c13\u4e16\u754c\u6a21\u578b\u5462\uff1f\u6309\u7167\u6700\u521dwayve\u5c55\u793a\u7684demo\uff0c\u4e16\u754c\u6a21\u578b\u4f9d\u8d56\u5b9e\u8f66\u91c7\u96c6\u7684\u6d77\u91cf\u6570\u636e\uff0c\u57fa\u4e8e\u751f\u6210\u6a21\u578b\u53bb\u751f\u6210\u672a\u6765\u573a\u666f\u6765\u548c\u771f\u5b9e\u7684\u672a\u6765\u65f6\u523b\u6570\u636e\uff0c\u8fdb\u800c\u8fdb\u884c\u76d1\u7763\uff0c\u8fd9\u662f\u5178\u578b\u7684\u65e0\u76d1\u7763\u8bad\u7ec3\u3002\u5176\u6700\u5de7\u5999\u7684\u5730\u65b9\u5219\u5728\u4e8e\u8981\u60f3\u6210\u529f\u9884\u6d4b\u672a\u6765\u65f6\u523b\u7684\u573a\u666f\uff0c\u4f60\u5fc5\u987b\u5bf9\u73b0\u5728\u65f6\u523b\u573a\u666f\u7684\u8bed\u4e49\u4fe1\u606f\u4ee5\u53ca\u4e16\u754c\u6f14\u5316\u7684\u89c4\u5f8b\u6709\u7740\u6df1\u523b\u7684\u4e86\u89e3\u3002\u5f53\u4e0b\u81ea\u52a8\u9a7e\u9a76\u65b9\u5411\u7684\u4e16\u754c\u6a21\u578b\u53ef\u4ee5\u5206\u6210\u4e24\u5927\u7c7b\uff1a\u751f\u6210\u5f0f\u548c\u7aef\u5230\u7aef\u3002\u4eca\u5929\u81ea\u52a8\u9a7e\u9a76\u4e4b\u5fc3\u5c31\u548c\u5927\u5bb6\u4e00\u8d77\u76d8\u70b9\u4e00\u4e0b\u4eca\u5e74\u4ee5\u6765\u8fd9\u65b9\u9762\u7684\u5de5\u4f5c\uff0c\u6587\u672b\u603b\u7ed3\uff01<\/p>\n

RenderWorld: World Model with Self-Supervised 3D Label<\/h3>\n
    \n
  • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2409.11356v1<\/li>\n<\/ul>\n

    \u4e0a\u6d77\u79d1\u6280\u5927\u5b66\u7684\u5de5\u4f5c\uff1a\u4ec5\u4f7f\u7528\u89c6\u89c9\u7684\u7aef\u5230\u7aef\u81ea\u52a8\u9a7e\u9a76\u4e0d\u4ec5\u6bd4LiDAR\u89c6\u89c9\u878d\u5408\u66f4\u5177\u6210\u672c\u6548\u76ca\uff0c\u800c\u4e14\u6bd4\u4f20\u7edf\u65b9\u6cd5\u66f4\u53ef\u9760\u3002\u4e3a\u4e86\u5b9e\u73b0\u7ecf\u6d4e\u4e14\u7a33\u5065\u7684\u7eaf\u89c6\u89c9\u81ea\u52a8\u9a7e\u9a76\u7cfb\u7edf\uff0c\u6211\u4eec\u63d0\u51fa\u4e86RenderWorld\uff0c\u8fd9\u662f\u4e00\u79cd\u4ec5\u652f\u6301\u89c6\u89c9\u7684\u7aef\u5230\u7aef\u81ea\u52a8\u9a7e\u9a76\u6846\u67b6\uff0c\u5b83\u4f7f\u7528\u57fa\u4e8e\u81ea\u76d1\u7763\u9ad8\u65af\u7684Img2Occ\u6a21\u5757\u751f\u62103D\u5360\u7528\u6807\u7b7e\uff0c\u7136\u540e\u901a\u8fc7AM-VAE\u5bf9\u6807\u7b7e\u8fdb\u884c\u7f16\u7801\uff0c\u5e76\u4f7f\u7528\u4e16\u754c\u6a21\u578b\u8fdb\u884c\u9884\u6d4b\u548c\u89c4\u5212\u3002RenderWorld\u91c7\u7528\u9ad8\u65af\u6563\u5c04\u6765\u8868\u793a3D\u573a\u666f\u548c\u6e32\u67d32D\u56fe\u50cf\uff0c\u4e0e\u57fa\u4e8eNeRF\u7684\u65b9\u6cd5\u76f8\u6bd4\uff0c\u5927\u5927\u63d0\u9ad8\u4e86\u5206\u5272\u7cbe\u5ea6\u5e76\u964d\u4f4e\u4e86GPU\u5185\u5b58\u6d88\u8017\u3002\u901a\u8fc7\u5e94\u7528AM-VAE\u5206\u522b\u5bf9\u7a7a\u6c14\u548c\u975e\u7a7a\u6c14\u8fdb\u884c\u7f16\u7801\uff0cRenderWorld\u5b9e\u73b0\u4e86\u66f4\u7ec6\u7c92\u5ea6\u7684\u573a\u666f\u5143\u7d20\u8868\u793a\uff0c\u4ece\u800c\u5728\u81ea\u56de\u5f52\u4e16\u754c\u6a21\u578b\u76844D\u5360\u7528\u9884\u6d4b\u548c\u8fd0\u52a8\u89c4\u5212\u65b9\u9762\u53d6\u5f97\u4e86\u6700\u5148\u8fdb\u7684\u6027\u80fd\u3002<\/p>\n

     <\/p>\n

    OccLLaMA: An Occupancy-Language-Action Generative World Model for Autonomous Driving<\/h3>\n
      \n
    • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2409.03272v1<\/li>\n<\/ul>\n

      \u590d\u65e6\u548c\u6e05\u534e\u7b49\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u591a\u6a21\u6001\u5927\u8bed\u8a00\u6a21\u578b\uff08MLLM\uff09\u7684\u5174\u8d77\u523a\u6fc0\u4e86\u5b83\u4eec\u5728\u81ea\u52a8\u9a7e\u9a76\u4e2d\u7684\u5e94\u7528\u3002\u6700\u8fd1\u57fa\u4e8eMLLM\u7684\u65b9\u6cd5\u901a\u8fc7\u5b66\u4e60\u4ece\u611f\u77e5\u5230\u884c\u52a8\u7684\u76f4\u63a5\u6620\u5c04\u6765\u5b9e\u73b0\u6700\u7ec8\u63a7\u5236\uff0c\u5ffd\u7565\u4e86\u4e16\u754c\u7684\u52a8\u6001\u4ee5\u53ca\u884c\u52a8\u4e0e\u4e16\u754c\u52a8\u6001\u4e4b\u95f4\u7684\u5173\u7cfb\u3002\u76f8\u6bd4\u4e4b\u4e0b\uff0c\u4eba\u7c7b\u62e5\u6709\u4e16\u754c\u6a21\u578b\uff0c\u4f7f\u4ed6\u4eec\u80fd\u591f\u57fa\u4e8e3D\u5185\u90e8\u89c6\u89c9\u8868\u793a\u6765\u6a21\u62df\u672a\u6765\u7684\u72b6\u6001\uff0c\u5e76\u76f8\u5e94\u5730\u8ba1\u5212\u884c\u52a8\u3002\u4e3a\u6b64\uff0c\u6211\u4eec\u63d0\u51fa\u4e86OccLLaMA\uff0c\u8fd9\u662f\u4e00\u79cd\u5360\u7528\u8bed\u8a00\u52a8\u4f5c\u751f\u6210\u4e16\u754c\u6a21\u578b\uff0c\u5b83\u4f7f\u7528\u8bed\u4e49\u5360\u7528\u4f5c\u4e3a\u4e00\u822c\u7684\u89c6\u89c9\u8868\u793a\uff0c\u5e76\u901a\u8fc7\u81ea\u56de\u5f52\u6a21\u578b\u7edf\u4e00\u89c6\u89c9\u8bed\u8a00\u52a8\u4f5c\uff08VLA\uff09\u6a21\u5f0f\u3002\u5177\u4f53\u6765\u8bf4\uff0c\u6211\u4eec\u5f15\u5165\u4e86\u4e00\u79cd\u65b0\u7684\u7c7b\u4f3cVQVAE\u7684\u573a\u666f\u6807\u8bb0\u5668\uff0c\u4ee5\u6709\u6548\u5730\u79bb\u6563\u548c\u91cd\u5efa\u8bed\u4e49\u5360\u7528\u573a\u666f\uff0c\u540c\u65f6\u8003\u8651\u5230\u5176\u7a00\u758f\u6027\u548c\u7c7b\u4e0d\u5e73\u8861\u6027\u3002\u7136\u540e\uff0c\u6211\u4eec\u4e3a\u89c6\u89c9\u3001\u8bed\u8a00\u548c\u52a8\u4f5c\u6784\u5efa\u4e86\u4e00\u4e2a\u7edf\u4e00\u7684\u591a\u6a21\u6001\u8bcd\u6c47\u8868\u3002\u6b64\u5916\uff0c\u6211\u4eec\u589e\u5f3a\u4e86LLM\uff0c\u7279\u522b\u662fLLaMA\uff0c\u4ee5\u5bf9\u7edf\u4e00\u8bcd\u6c47\u8868\u6267\u884c\u4e0b\u4e00\u4e2a\u4ee4\u724c\/\u573a\u666f\u9884\u6d4b\uff0c\u4ece\u800c\u5b8c\u6210\u81ea\u52a8\u9a7e\u9a76\u4e2d\u7684\u591a\u9879\u4efb\u52a1\u3002\u5927\u91cf\u5b9e\u9a8c\u8868\u660e\uff0cOccLLaMA\u5728\u591a\u4e2a\u4efb\u52a1\u4e2d\u90fd\u53d6\u5f97\u4e86\u5177\u6709\u7ade\u4e89\u529b\u7684\u6027\u80fd\uff0c\u5305\u62ec4D\u5360\u7528\u9884\u6d4b\u3001\u8fd0\u52a8\u89c4\u5212\u548c\u89c6\u89c9\u95ee\u7b54\uff0c\u5c55\u793a\u4e86\u5176\u4f5c\u4e3a\u81ea\u52a8\u9a7e\u9a76\u57fa\u7840\u6a21\u578b\u7684\u6f5c\u529b\u3002<\/p>\n

       <\/p>\n

      Mitigating Covariate Shift in Imitation Learning for Autonomous Vehicles Using Latent Space Generative World Models<\/h3>\n
        \n
      • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2409.16663v2<\/li>\n<\/ul>\n

        \u82f1\u4f1f\u8fbe\u7684\u5de5\u4f5c\uff1a\u6211\u4eec\u5efa\u8bae\u4f7f\u7528\u6f5c\u5728\u7a7a\u95f4\u751f\u6210\u4e16\u754c\u6a21\u578b\u6765\u89e3\u51b3\u81ea\u52a8\u9a7e\u9a76\u4e2d\u7684\u534f\u53d8\u91cf\u8f6c\u6362\u95ee\u9898\u3002\u4e16\u754c\u6a21\u578b\u662f\u4e00\u79cd\u795e\u7ecf\u7f51\u7edc\uff0c\u80fd\u591f\u6839\u636e\u8fc7\u53bb\u7684\u72b6\u6001\u548c\u52a8\u4f5c\u9884\u6d4b\u4ee3\u7406\u7684\u4e0b\u4e00\u4e2a\u72b6\u6001\u3002\u901a\u8fc7\u5728\u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u5229\u7528\u4e16\u754c\u6a21\u578b\uff0c\u9a7e\u9a76\u7b56\u7565\u6709\u6548\u5730\u7f13\u89e3\u4e86\u534f\u53d8\u91cf\u53d8\u5316\uff0c\u800c\u4e0d\u9700\u8981\u8fc7\u591a\u7684\u8bad\u7ec3\u6570\u636e\u3002\u5728\u7aef\u5230\u7aef\u8bad\u7ec3\u671f\u95f4\uff0c\u6211\u4eec\u7684\u7b56\u7565\u901a\u8fc7\u4e0e\u4eba\u7c7b\u6f14\u793a\u4e2d\u89c2\u5bdf\u5230\u7684\u72b6\u6001\u5bf9\u9f50\u6765\u5b66\u4e60\u5982\u4f55\u4ece\u9519\u8bef\u4e2d\u6062\u590d\uff0c\u4ee5\u4fbf\u5728\u8fd0\u884c\u65f6\u53ef\u4ee5\u4ece\u8bad\u7ec3\u5206\u5e03\u4e4b\u5916\u7684\u6270\u52a8\u4e2d\u6062\u590d\u3002\u6b64\u5916\u6211\u4eec\u4ecb\u7ecd\u4e86\u4e00\u79cd\u57fa\u4e8eTransformer\u7684\u611f\u77e5\u7f16\u7801\u5668\uff0c\u8be5\u7f16\u7801\u5668\u91c7\u7528\u591a\u89c6\u56fe\u4ea4\u53c9\u6ce8\u610f\u529b\u548c\u5b66\u4e60\u573a\u666f\u67e5\u8be2\u3002\u6211\u4eec\u5448\u73b0\u4e86\u5b9a\u6027\u548c\u5b9a\u91cf\u7ed3\u679c\uff0c\u5c55\u793a\u4e86\u5728CARLA\u6a21\u62df\u5668\u95ed\u73af\u6d4b\u8bd5\u65b9\u9762\u5bf9\u73b0\u6709\u6280\u672f\u7684\u663e\u8457\u6539\u8fdb\uff0c\u5e76\u5c55\u793a\u4e86CARLA\u548cNVIDIA DRIVE Sim\u5904\u7406\u6270\u52a8\u7684\u80fd\u529b\u3002<\/p>\n

         <\/p>\n

        Driving in the Occupancy World: Vision-Centric 4D Occupancy Forecasting and Planning via World Models for Autonomous Driving<\/h3>\n
          \n
        • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2408.14197v1<\/li>\n<\/ul>\n

          \u6d59\u5927&\u534e\u4e3a\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u4e16\u754c\u6a21\u578b\u57fa\u4e8e\u5404\u79cd\u81ea\u8f66\u884c\u4e3a\u8bbe\u60f3\u4e86\u6f5c\u5728\u7684\u672a\u6765\u72b6\u6001\u3002\u5b83\u4eec\u5d4c\u5165\u4e86\u5173\u4e8e\u9a7e\u9a76\u73af\u5883\u7684\u5e7f\u6cdb\u77e5\u8bc6\uff0c\u4fc3\u8fdb\u4e86\u5b89\u5168\u548c\u53ef\u6269\u5c55\u7684\u81ea\u52a8\u9a7e\u9a76\u3002\u5927\u591a\u6570\u73b0\u6709\u65b9\u6cd5\u4e3b\u8981\u5173\u6ce8\u6570\u636e\u751f\u6210\u6216\u4e16\u754c\u6a21\u578b\u7684\u9884\u8bad\u7ec3\u8303\u5f0f\u3002\u4e0e\u4e0a\u8ff0\u5148\u524d\u7684\u5de5\u4f5c\u4e0d\u540c\uff0c\u6211\u4eec\u63d0\u51fa\u4e86Drive OccWorld\uff0c\u5b83\u5c06\u4ee5\u89c6\u89c9\u4e3a\u4e2d\u5fc3\u76844D\u9884\u6d4b\u4e16\u754c\u6a21\u578b\u5e94\u7528\u4e8e\u81ea\u52a8\u9a7e\u9a76\u7684\u7aef\u5230\u7aef\u89c4\u5212\u3002\u5177\u4f53\u6765\u8bf4\uff0c\u6211\u4eec\u9996\u5148\u5728\u5185\u5b58\u6a21\u5757\u4e2d\u5f15\u5165\u8bed\u4e49\u548c\u8fd0\u52a8\u6761\u4ef6\u89c4\u8303\u5316\uff0c\u8be5\u6a21\u5757\u4ece\u5386\u53f2BEV\u5d4c\u5165\u4e2d\u79ef\u7d2f\u8bed\u4e49\u548c\u52a8\u6001\u4fe1\u606f\u3002\u7136\u540e\u5c06\u8fd9\u4e9bBEV\u7279\u5f81\u4f20\u9001\u5230\u4e16\u754c\u89e3\u7801\u5668\uff0c\u4ee5\u8fdb\u884c\u672a\u6765\u7684\u5360\u7528\u548c\u6d41\u91cf\u9884\u6d4b\uff0c\u540c\u65f6\u8003\u8651\u51e0\u4f55\u548c\u65f6\u7a7a\u5efa\u6a21\u3002\u6b64\u5916\uff0c\u6211\u4eec\u5efa\u8bae\u5728\u4e16\u754c\u6a21\u578b\u4e2d\u6ce8\u5165\u7075\u6d3b\u7684\u52a8\u4f5c\u6761\u4ef6\uff0c\u5982\u901f\u5ea6\u3001\u8f6c\u5411\u89d2\u3001\u8f68\u8ff9\u548c\u547d\u4ee4\uff0c\u4ee5\u5b9e\u73b0\u53ef\u63a7\u751f\u6210\uff0c\u5e76\u4fc3\u8fdb\u66f4\u5e7f\u6cdb\u7684\u4e0b\u6e38\u5e94\u7528\u3002\u6b64\u5916\uff0c\u6211\u4eec\u63a2\u7d22\u5c064D\u4e16\u754c\u6a21\u578b\u7684\u751f\u6210\u80fd\u529b\u4e0e\u7aef\u5230\u7aef\u89c4\u5212\u76f8\u7ed3\u5408\uff0c\u4ece\u800c\u80fd\u591f\u4f7f\u7528\u57fa\u4e8e\u5360\u7528\u7684\u6210\u672c\u51fd\u6570\u5bf9\u672a\u6765\u72b6\u6001\u8fdb\u884c\u8fde\u7eed\u9884\u6d4b\u5e76\u9009\u62e9\u6700\u4f73\u8f68\u8ff9\u3002\u5bf9nuScenes\u6570\u636e\u96c6\u7684\u5e7f\u6cdb\u5b9e\u9a8c\u8868\u660e\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u53ef\u4ee5\u751f\u6210\u5408\u7406\u53ef\u63a7\u76844D\u5360\u7528\u7387\uff0c\u4e3a\u63a8\u52a8\u4e16\u754c\u751f\u6210\u548c\u7aef\u5230\u7aef\u89c4\u5212\u5f00\u8f9f\u4e86\u65b0\u9014\u5f84\u3002<\/p>\n

           <\/p>\n

          BEVWorld: A Multimodal World Model for Autonomous Driving via Unified BEV Latent Space<\/h3>\n
            \n
          • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2407.05679v2<\/li>\n
          • \u5f00\u6e90\u94fe\u63a5\uff1ahttps:\/\/github.com\/zympsyche\/BevWorld<\/li>\n<\/ul>\n

            \u767e\u5ea6\u7684\u5de5\u4f5c\uff1a\u4e16\u754c\u6a21\u578b\u56e0\u5176\u9884\u6d4b\u6f5c\u5728\u672a\u6765\u60c5\u666f\u7684\u80fd\u529b\u800c\u5728\u81ea\u52a8\u9a7e\u9a76\u9886\u57df\u53d7\u5230\u8d8a\u6765\u8d8a\u591a\u7684\u5173\u6ce8\u3002\u5728\u672c\u6587\u4e2d\uff0c\u6211\u4eec\u63d0\u51fa\u4e86BEVWorld\uff0c\u8fd9\u662f\u4e00\u79cd\u5c06\u591a\u6a21\u6001\u4f20\u611f\u5668\u8f93\u5165\u6807\u8bb0\u4e3a\u7edf\u4e00\u7d27\u51d1\u7684\u9e1f\u77b0\u56fe\uff08BEV\uff09\u6f5c\u5728\u7a7a\u95f4\u4ee5\u8fdb\u884c\u73af\u5883\u5efa\u6a21\u7684\u65b0\u65b9\u6cd5\u3002\u4e16\u754c\u6a21\u578b\u7531\u4e24\u90e8\u5206\u7ec4\u6210\uff1a\u591a\u6a21\u6001\u6807\u8bb0\u5668\u548c\u6f5c\u5728BEV\u5e8f\u5217\u6269\u6563\u6a21\u578b\u3002\u591a\u6a21\u6001\u6807\u8bb0\u5668\u9996\u5148\u5bf9\u591a\u6a21\u6001\u4fe1\u606f\u8fdb\u884c\u7f16\u7801\uff0c\u89e3\u7801\u5668\u80fd\u591f\u4ee5\u81ea\u76d1\u7763\u7684\u65b9\u5f0f\u901a\u8fc7\u5149\u7ebf\u6295\u5c04\u6e32\u67d3\u5c06\u6f5c\u5728\u7684BEV\u6807\u8bb0\u91cd\u5efa\u4e3aLiDAR\u548c\u56fe\u50cf\u89c2\u6d4b\u3002\u7136\u540e\uff0c\u6f5c\u5728\u7684BEV\u5e8f\u5217\u6269\u6563\u6a21\u578b\u5728\u7ed9\u5b9a\u52a8\u4f5c\u6807\u8bb0\u4f5c\u4e3a\u6761\u4ef6\u7684\u60c5\u51b5\u4e0b\u9884\u6d4b\u672a\u6765\u7684\u60c5\u666f\u3002\u5b9e\u9a8c\u8bc1\u660e\u4e86BEVWorld\u5728\u81ea\u52a8\u9a7e\u9a76\u4efb\u52a1\u4e2d\u7684\u6709\u6548\u6027\uff0c\u5c55\u793a\u4e86\u5176\u751f\u6210\u672a\u6765\u573a\u666f\u7684\u80fd\u529b\uff0c\u5e76\u4f7f\u611f\u77e5\u548c\u8fd0\u52a8\u9884\u6d4b\u7b49\u4e0b\u6e38\u4efb\u52a1\u53d7\u76ca\u3002<\/p>\n

             <\/p>\n

            Planning with Adaptive World Models for Autonomous Driving<\/h3>\n
              \n
            • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2406.10714v2<\/li>\n
            • \u9879\u76ee\u4e3b\u9875\uff1ahttps:\/\/arunbalajeev.github.io\/world_models_planning\/world_model_paper.html<\/li>\n<\/ul>\n

              \u5361\u5185\u57fa\u6885\u9686\u5927\u5b66\u7684\u5de5\u4f5c\uff1a\u8fd0\u52a8\u89c4\u5212\u5bf9\u4e8e\u590d\u6742\u57ce\u5e02\u73af\u5883\u4e2d\u7684\u5b89\u5168\u5bfc\u822a\u81f3\u5173\u91cd\u8981\u3002\u4ece\u5386\u53f2\u4e0a\u770b\uff0c\u8fd0\u52a8\u89c4\u5212\u5668\uff08MP\uff09\u5df2\u7ecf\u7528\u7a0b\u5e8f\u751f\u6210\u7684\u6a21\u62df\u5668\uff08\u5982CARLA\uff09\u8fdb\u884c\u4e86\u8bc4\u4f30\u3002\u7136\u800c\uff0c\u8fd9\u79cd\u5408\u6210\u57fa\u51c6\u5e76\u4e0d\u80fd\u6355\u6349\u5230\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u591a\u667a\u80fd\u4f53\u4ea4\u4e92\u3002nuPlan\u662f\u6700\u8fd1\u53d1\u5e03\u7684MP\u57fa\u51c6\u6d4b\u8bd5\uff0c\u5b83\u901a\u8fc7\u7528\u95ed\u73af\u4eff\u771f\u903b\u8f91\u589e\u5f3a\u73b0\u5b9e\u4e16\u754c\u7684\u9a7e\u9a76\u65e5\u5fd7\u6765\u89e3\u51b3\u8fd9\u4e00\u5c40\u9650\u6027\uff0c\u6709\u6548\u5730\u5c06\u56fa\u5b9a\u6570\u636e\u96c6\u8f6c\u5316\u4e3a\u53cd\u5e94\u5f0f\u6a21\u62df\u5668\u3002\u6211\u4eec\u5206\u6790\u4e86nuPlan\u8bb0\u5f55\u65e5\u5fd7\u7684\u7279\u5f81\uff0c\u53d1\u73b0\u6bcf\u4e2a\u57ce\u5e02\u90fd\u6709\u81ea\u5df1\u72ec\u7279\u7684\u9a7e\u9a76\u884c\u4e3a\uff0c\u8fd9\u8868\u660e\u7a33\u5065\u7684\u89c4\u5212\u8005\u5fc5\u987b\u9002\u5e94\u4e0d\u540c\u7684\u73af\u5883\u3002\u6211\u4eec\u5b66\u4e60\u4f7f\u7528BehaviorNet\u5bf9\u8fd9\u79cd\u72ec\u7279\u7684\u884c\u4e3a\u8fdb\u884c\u5efa\u6a21\uff0cBehaviorNet\u662f\u4e00\u79cd\u56fe\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff08GCNN\uff09\uff0c\u5b83\u4f7f\u7528\u6700\u8fd1\u89c2\u5bdf\u5230\u7684\u4ee3\u7406\u5386\u53f2\u4e2d\u5f97\u51fa\u7684\u7279\u5f81\u6765\u9884\u6d4b\u53cd\u5e94\u6027\u4ee3\u7406\u884c\u4e3a\uff1b\u76f4\u89c9\u4e0a\uff0c\u4e00\u4e9b\u6fc0\u8fdb\u7684\u7279\u5de5\u53ef\u80fd\u4f1a\u5c3e\u968f\u9886\u5148\u7684\u8f66\u8f86\uff0c\u800c\u53e6\u4e00\u4e9b\u5219\u53ef\u80fd\u4e0d\u4f1a\u3002\u4e3a\u4e86\u6a21\u62df\u8fd9\u79cd\u73b0\u8c61\uff0cBehaviorNet\u9884\u6d4b\u4ee3\u7406\u8fd0\u52a8\u63a7\u5236\u5668\u7684\u53c2\u6570\uff0c\u800c\u4e0d\u662f\u76f4\u63a5\u9884\u6d4b\u5176\u65f6\u7a7a\u8f68\u8ff9\uff08\u5c31\u50cf\u5927\u591a\u6570\u9884\u6d4b\u8005\u90a3\u6837\uff09\u3002\u6700\u540e\uff0c\u6211\u4eec\u63d0\u51fa\u4e86AdaptiveDriver\uff0c\u8fd9\u662f\u4e00\u79cd\u57fa\u4e8e\u6a21\u578b\u9884\u6d4b\u63a7\u5236\uff08MPC\uff09\u7684\u89c4\u5212\u5668\uff0c\u53ef\u4ee5\u5c55\u5f00\u57fa\u4e8eBehaviorNet\u9884\u6d4b\u7684\u4e0d\u540c\u4e16\u754c\u6a21\u578b\u3002\u6211\u4eec\u5e7f\u6cdb\u7684\u5b9e\u9a8c\u8868\u660e\uff0cAdaptiveDriver\u5728nuPlan\u95ed\u73af\u89c4\u5212\u57fa\u51c6\u4e0a\u53d6\u5f97\u4e86\u6700\u5148\u8fdb\u7684\u7ed3\u679c\uff0c\u5728Test-14 Hard R-CLS\u4e0a\u6bd4\u4e4b\u524d\u7684\u5de5\u4f5c\u63d0\u9ad8\u4e862%\uff0c\u5373\u4f7f\u5728\u4ece\u672a\u89c1\u8fc7\u7684\u57ce\u5e02\u8fdb\u884c\u8bc4\u4f30\u65f6\u4e5f\u5177\u6709\u666e\u904d\u6027\u3002<\/p>\n

               <\/p>\n

              Enhancing End-to-End Autonomous Driving with Latent World Model<\/h3>\n
                \n
              • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2406.08481v1<\/li>\n<\/ul>\n

                \u4e2d\u79d1\u9662\u548c\u4e2d\u79d1\u9662\u81ea\u52a8\u5316\u7814\u7a76\u6240\u7b49\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u7aef\u5230\u7aef\u81ea\u52a8\u9a7e\u9a76\u5f15\u8d77\u4e86\u5e7f\u6cdb\u5173\u6ce8\u3002\u5f53\u524d\u7684\u7aef\u5230\u7aef\u65b9\u6cd5\u5728\u5f88\u5927\u7a0b\u5ea6\u4e0a\u4f9d\u8d56\u4e8e\u611f\u77e5\u4efb\u52a1\u7684\u76d1\u7763\uff0c\u5982\u68c0\u6d4b\u3001\u8ddf\u8e2a\u548c\u5730\u56fe\u5206\u5272\uff0c\u4ee5\u5e2e\u52a9\u5b66\u4e60\u573a\u666f\u8868\u793a\u3002\u7136\u800c\uff0c\u8fd9\u4e9b\u65b9\u6cd5\u9700\u8981\u5927\u91cf\u7684\u6807\u6ce8\uff0c\u963b\u788d\u4e86\u6570\u636e\u7684\u53ef\u6269\u5c55\u6027\u3002\u4e3a\u4e86\u5e94\u5bf9\u8fd9\u4e00\u6311\u6218\uff0c\u6211\u4eec\u63d0\u51fa\u4e86\u4e00\u79cd\u65b0\u7684\u81ea\u76d1\u7763\u65b9\u6cd5\u6765\u589e\u5f3a\u7aef\u5230\u7aef\u7684\u9a71\u52a8\uff0c\u800c\u4e0d\u9700\u8981\u6602\u8d35\u7684\u6807\u7b7e\u3002\u5177\u4f53\u6765\u8bf4\uff0c\u6211\u4eec\u7684\u6846\u67b6LAW\u4f7f\u7528LAtent World model\uff0c\u6839\u636e\u9884\u6d4b\u7684\u81ea\u8f66\u884c\u4e3a\u548c\u5f53\u524d\u6846\u67b6\u7684\u6f5c\u5728\u7279\u5f81\u6765\u9884\u6d4b\u672a\u6765\u7684\u6f5c\u5728\u7279\u5f81\u3002\u9884\u6d4b\u7684\u6f5c\u5728\u7279\u5f81\u7531\u672a\u6765\u5b9e\u9645\u89c2\u5bdf\u5230\u7684\u7279\u5f81\u8fdb\u884c\u76d1\u7763\u3002\u8fd9\u79cd\u76d1\u7763\u8054\u5408\u4f18\u5316\u4e86\u6f5c\u5728\u7279\u5f81\u5b66\u4e60\u548c\u52a8\u4f5c\u9884\u6d4b\uff0c\u5927\u5927\u63d0\u9ad8\u4e86\u9a7e\u9a76\u6027\u80fd\u3002\u56e0\u6b64\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u5728\u5f00\u73af\u548c\u95ed\u73af\u57fa\u51c6\u6d4b\u8bd5\u4e2d\u90fd\u5b9e\u73b0\u4e86\u6700\u5148\u8fdb\u7684\u6027\u80fd\uff0c\u800c\u65e0\u9700\u6602\u8d35\u7684\u6807\u6ce8\u3002<\/p>\n

                 <\/p>\n

                Probing Multimodal LLMs as World Models for Driving<\/h3>\n
                  \n
                • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2405.05956v1<\/li>\n
                • \u5f00\u6e90\u94fe\u63a5\uff1ahttps:\/\/github.com\/sreeramsa\/DriveSim<\/li>\n<\/ul>\n

                  MIT\u7b49\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u6211\u4eec\u51b7\u9759\u5730\u770b\u5f85\u4e86\u591a\u6a21\u6001\u5927\u8bed\u8a00\u6a21\u578b\uff08MLLM\uff09\u5728\u81ea\u52a8\u9a7e\u9a76\u9886\u57df\u7684\u5e94\u7528\uff0c\u5e76\u6311\u6218\/\u9a8c\u8bc1\u4e86\u4e00\u4e9b\u5e38\u89c1\u7684\u5047\u8bbe\uff0c\u91cd\u70b9\u662f\u5b83\u4eec\u5728\u95ed\u73af\u63a7\u5236\u73af\u5883\u4e2d\u901a\u8fc7\u56fe\u50cf\/\u5e27\u5e8f\u5217\u63a8\u7406\u548c\u89e3\u91ca\u52a8\u6001\u9a7e\u9a76\u573a\u666f\u7684\u80fd\u529b\u3002\u5c3d\u7ba1GPT-4V\u7b49MLLM\u53d6\u5f97\u4e86\u91cd\u5927\u8fdb\u5c55\uff0c\u4f46\u5b83\u4eec\u5728\u590d\u6742\u3001\u52a8\u6001\u9a7e\u9a76\u73af\u5883\u4e2d\u7684\u6027\u80fd\u5728\u5f88\u5927\u7a0b\u5ea6\u4e0a\u4ecd\u672a\u7ecf\u8fc7\u6d4b\u8bd5\uff0c\u8fd9\u662f\u4e00\u4e2a\u5e7f\u6cdb\u7684\u63a2\u7d22\u9886\u57df\u3002\u6211\u4eec\u8fdb\u884c\u4e86\u4e00\u9879\u5168\u9762\u7684\u5b9e\u9a8c\u7814\u7a76\uff0c\u4ece\u56fa\u5b9a\u8f66\u8f7d\u6444\u50cf\u5934\u7684\u89d2\u5ea6\u8bc4\u4f30\u5404\u79cdMLLM\u4f5c\u4e3a\u4e16\u754c\u9a7e\u9a76\u6a21\u578b\u7684\u80fd\u529b\u3002\u6211\u4eec\u7684\u7814\u7a76\u7ed3\u679c\u8868\u660e\uff0c\u867d\u7136\u8fd9\u4e9b\u6a21\u578b\u80fd\u591f\u719f\u7ec3\u5730\u89e3\u91ca\u5355\u4e2a\u56fe\u50cf\uff0c\u4f46\u5b83\u4eec\u5728\u8de8\u63cf\u8ff0\u52a8\u6001\u884c\u4e3a\u7684\u6846\u67b6\u5408\u6210\u8fde\u8d2f\u7684\u53d9\u4e8b\u6216\u903b\u8f91\u5e8f\u5217\u65b9\u9762\u5b58\u5728\u5f88\u5927\u56f0\u96be\u3002\u5b9e\u9a8c\u8868\u660e\uff0c\u5728\u9884\u6d4b\uff08i\uff09\u57fa\u672c\u8f66\u8f86\u52a8\u529b\u5b66\uff08\u524d\u8fdb\/\u540e\u9000\u3001\u52a0\u901f\/\u51cf\u901f\u3001\u53f3\u8f6c\u6216\u5de6\u8f6c\uff09\u3001\uff08ii\uff09\u4e0e\u5176\u4ed6\u9053\u8def\u53c2\u4e0e\u8005\u7684\u76f8\u4e92\u4f5c\u7528\uff08\u4f8b\u5982\uff0c\u8bc6\u522b\u8d85\u901f\u884c\u9a76\u7684\u6c7d\u8f66\u6216\u7e41\u5fd9\u7684\u4ea4\u901a\uff09\u3001\uff08iii\uff09\u8f68\u8ff9\u89c4\u5212\u548c\uff08iv\uff09\u5f00\u653e\u96c6\u52a8\u6001\u573a\u666f\u63a8\u7406\u65b9\u9762\u5b58\u5728\u76f8\u5f53\u5927\u7684\u4e0d\u51c6\u786e\u6027\uff0c\u8fd9\u8868\u660e\u6a21\u578b\u8bad\u7ec3\u6570\u636e\u4e2d\u5b58\u5728\u504f\u5dee\u3002\u4e3a\u4e86\u5b9e\u73b0\u8fd9\u9879\u5b9e\u9a8c\u7814\u7a76\uff0c\u6211\u4eec\u5f15\u5165\u4e86\u4e00\u4e2a\u4e13\u95e8\u7684\u6a21\u62df\u5668DriveSim\uff0c\u65e8\u5728\u751f\u6210\u5404\u79cd\u9a7e\u9a76\u573a\u666f\uff0c\u4e3a\u8bc4\u4f30\u9a7e\u9a76\u9886\u57df\u7684MLLM\u63d0\u4f9b\u5e73\u53f0\u3002\u6b64\u5916\uff0c\u6211\u4eec\u8fd8\u8d21\u732e\u4e86\u5b8c\u6574\u7684\u5f00\u6e90\u4ee3\u7801\u548c\u4e00\u4e2a\u65b0\u7684\u6570\u636e\u96c6\u201cEval LLM Drive\u201d\uff0c\u7528\u4e8e\u8bc4\u4f30\u9a7e\u9a76\u4e2d\u7684MLLM\u3002\u6211\u4eec\u7684\u7814\u7a76\u7ed3\u679c\u7a81\u663e\u4e86\u5f53\u524d\u6700\u5148\u8fdbMLLM\u80fd\u529b\u7684\u4e00\u4e2a\u5173\u952e\u5dee\u8ddd\uff0c\u5f3a\u8c03\u4e86\u589e\u5f3a\u57fa\u7840\u6a21\u578b\u7684\u5fc5\u8981\u6027\uff0c\u4ee5\u63d0\u9ad8\u5176\u5728\u73b0\u5b9e\u4e16\u754c\u52a8\u6001\u73af\u5883\u4e2d\u7684\u9002\u7528\u6027\u3002<\/p>\n

                   <\/p>\n

                  OccSora: 4D Occupancy Generation Models as World Simulators for Autonomous Driving<\/h3>\n
                    \n
                  • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2405.20337<\/li>\n
                  • \u5f00\u6e90\u94fe\u63a5\uff1ahttps:\/\/github.com\/wzzheng\/OccSora<\/li>\n<\/ul>\n

                    \u5317\u822a&UC Berkeley\u7b49\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u4e86\u89e33D\u573a\u666f\u7684\u6f14\u53d8\u5bf9\u4e8e\u6709\u6548\u7684\u81ea\u52a8\u9a7e\u9a76\u975e\u5e38\u91cd\u8981\u3002\u867d\u7136\u4f20\u7edf\u65b9\u6cd5\u5c06\u573a\u666f\u5f00\u53d1\u4e0e\u5355\u4e2a\u5b9e\u4f8b\u7684\u8fd0\u52a8\u76f8\u7ed3\u5408\uff0c\u4f46\u4e16\u754c\u6a21\u578b\u4f5c\u4e3a\u4e00\u4e2a\u751f\u6210\u6846\u67b6\u51fa\u73b0\uff0c\u7528\u4e8e\u63cf\u8ff0\u4e00\u822c\u7684\u573a\u666f\u52a8\u6001\u3002\u7136\u800c\u5927\u591a\u6570\u73b0\u6709\u65b9\u6cd5\u91c7\u7528\u81ea\u56de\u5f52\u6846\u67b6\u6765\u6267\u884c\u4e0b\u4e00\u4e2a\u4ee4\u724c\u9884\u6d4b\uff0c\u8fd9\u5728\u5efa\u6a21\u957f\u671f\u65f6\u95f4\u6f14\u5316\u65b9\u9762\u6548\u7387\u4f4e\u4e0b\u3002\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\uff0c\u6211\u4eec\u63d0\u51fa\u4e86\u4e00\u79cd\u57fa\u4e8e\u6269\u6563\u76844D\u5360\u7528\u751f\u6210\u6a21\u578bOccSora\uff0c\u6765\u6a21\u62df\u81ea\u52a8\u9a7e\u9a763D\u4e16\u754c\u7684\u53d1\u5c55\u3002\u6211\u4eec\u91c7\u75284D\u573a\u666f\u6807\u8bb0\u5668\u6765\u83b7\u5f974D\u5360\u7528\u8f93\u5165\u7684\u7d27\u51d1\u79bb\u6563\u65f6\u7a7a\u8868\u793a\uff0c\u5e76\u5b9e\u73b0\u957f\u5e8f\u5217\u5360\u7528\u89c6\u9891\u7684\u9ad8\u8d28\u91cf\u91cd\u5efa\u3002\u7136\u540e\uff0c\u6211\u4eec\u5b66\u4e60\u65f6\u7a7a\u8868\u793a\u4e0a\u7684\u6269\u6563Transformer\uff0c\u5e76\u6839\u636e\u8f68\u8ff9\u63d0\u793a\u751f\u62104D\u5360\u7528\u7387\u3002\u6211\u4eec\u5bf9\u5e7f\u6cdb\u4f7f\u7528\u7684\u5177\u6709Occ3D\u5360\u7528\u6ce8\u91ca\u7684nuScenes\u6570\u636e\u96c6\u8fdb\u884c\u4e86\u5e7f\u6cdb\u7684\u5b9e\u9a8c\u3002OccSora\u53ef\u4ee5\u751f\u6210\u5177\u6709\u771f\u5b9e3D\u5e03\u5c40\u548c\u65f6\u95f4\u4e00\u81f4\u6027\u768416\u79d2\u89c6\u9891\uff0c\u5c55\u793a\u4e86\u5176\u7406\u89e3\u9a7e\u9a76\u573a\u666f\u7684\u7a7a\u95f4\u548c\u65f6\u95f4\u5206\u5e03\u7684\u80fd\u529b\u3002\u901a\u8fc7\u8f68\u8ff9\u611f\u77e54D\u751f\u6210\uff0cOccSora\u6709\u53ef\u80fd\u6210\u4e3a\u81ea\u52a8\u9a7e\u9a76\u51b3\u7b56\u7684\u4e16\u754c\u6a21\u62df\u5668\u3002<\/p>\n

                     <\/p>\n

                    DriveDreamer-2: LLM-Enhanced World Models for Diverse Driving Video Generation<\/h3>\n
                      \n
                    • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2403.06845v2<\/li>\n
                    • \u9879\u76ee\u4e3b\u9875\uff1ahttps:\/\/drivedreamer2.github.io\/<\/li>\n<\/ul>\n

                      \u4e2d\u79d1\u9662\u81ea\u52a8\u5316\u7814\u7a76\u6240&GigaAI\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u4e16\u754c\u6a21\u578b\u5728\u81ea\u52a8\u9a7e\u9a76\u65b9\u9762\u8868\u73b0\u51fa\u4e86\u4f18\u52bf\uff0c\u7279\u522b\u662f\u5728\u751f\u6210\u591a\u89c6\u56fe\u9a7e\u9a76\u89c6\u9891\u65b9\u9762\u3002\u7136\u800c\uff0c\u5728\u751f\u6210\u5b9a\u5236\u7684\u9a7e\u9a76\u89c6\u9891\u65b9\u9762\u4ecd\u7136\u5b58\u5728\u91cd\u5927\u6311\u6218\u3002\u5728\u672c\u6587\u4e2d\uff0c\u6211\u4eec\u63d0\u51fa\u4e86DriveDreamer-2\uff0c\u5b83\u57fa\u4e8eDriveDreamer\u7684\u6846\u67b6\uff0c\u5e76\u7ed3\u5408\u4e86\u4e00\u4e2a\u5927\u578b\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09\u6765\u751f\u6210\u7528\u6237\u5b9a\u4e49\u7684\u9a7e\u9a76\u89c6\u9891\u3002\u5177\u4f53\u6765\u8bf4\uff0c\u6700\u521d\u7ed3\u5408\u4e86LLM\u63a5\u53e3\uff0c\u5c06\u7528\u6237\u7684\u67e5\u8be2\u8f6c\u6362\u4e3a\u4ee3\u7406\u8f68\u8ff9\u3002\u968f\u540e\uff0c\u6839\u636e\u8f68\u8ff9\u751f\u6210\u7b26\u5408\u4ea4\u901a\u89c4\u5219\u7684HDMap\u3002\u6700\u7ec8\uff0c\u6211\u4eec\u63d0\u51fa\u4e86\u7edf\u4e00\u591a\u89c6\u56fe\u6a21\u578b\u6765\u589e\u5f3a\u751f\u6210\u7684\u9a7e\u9a76\u89c6\u9891\u4e2d\u7684\u65f6\u95f4\u548c\u7a7a\u95f4\u8fde\u8d2f\u6027\u3002DriveDreamer-2\u662f\u4e16\u754c\u4e0a\u7b2c\u4e00\u6b3e\u751f\u6210\u5b9a\u5236\u9a7e\u9a76\u89c6\u9891\u7684\u8f66\u578b\uff0c\u5b83\u53ef\u4ee5\u4ee5\u7528\u6237\u53cb\u597d\u7684\u65b9\u5f0f\u751f\u6210\u4e0d\u5e38\u89c1\u7684\u9a7e\u9a76\u89c6\u9891\uff08\u4f8b\u5982\uff0c\u7a81\u7136\u5207\u5165\u7684\u8f66\u8f86\uff09\u3002\u6b64\u5916\uff0c\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\uff0c\u751f\u6210\u7684\u89c6\u9891\u589e\u5f3a\u4e86\u9a7e\u9a76\u611f\u77e5\u65b9\u6cd5\uff08\u59823D\u68c0\u6d4b\u548c\u8ddf\u8e2a\uff09\u7684\u8bad\u7ec3\u3002\u6b64\u5916\uff0cDriveDreamer-2\u7684\u89c6\u9891\u751f\u6210\u8d28\u91cf\u8d85\u8d8a\u4e86\u5176\u4ed6\u6700\u5148\u8fdb\u7684\u65b9\u6cd5\uff0c\u663e\u793aFID\u548cFVD\u5f97\u5206\u5206\u522b\u4e3a11.2\u548c55.7\uff0c\u76f8\u5bf9\u63d0\u9ad8\u4e8630%\u548c50%\u3002<\/p>\n

                       <\/p>\n

                      WorldDreamer: Towards General World Models for Video Generation via Predicting Masked Tokens<\/h3>\n
                        \n
                      • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2401.09985v1<\/li>\n
                      • \u9879\u76ee\u4e3b\u9875\uff1ahttps:\/\/world-dreamer.github.io\/<\/li>\n<\/ul>\n

                        GigaAI\u548c\u6e05\u534e\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u4e16\u754c\u6a21\u578b\u5728\u7406\u89e3\u548c\u9884\u6d4b\u4e16\u754c\u52a8\u6001\u65b9\u9762\u53d1\u6325\u7740\u81f3\u5173\u91cd\u8981\u7684\u4f5c\u7528\uff0c\u8fd9\u5bf9\u89c6\u9891\u751f\u6210\u81f3\u5173\u91cd\u8981\u3002\u7136\u800c\uff0c\u73b0\u6709\u7684\u4e16\u754c\u6a21\u578b\u4ec5\u9650\u4e8e\u6e38\u620f\u6216\u9a7e\u9a76\u7b49\u7279\u5b9a\u573a\u666f\uff0c\u9650\u5236\u4e86\u5b83\u4eec\u6355\u6349\u4e00\u822c\u4e16\u754c\u52a8\u6001\u73af\u5883\u590d\u6742\u6027\u7684\u80fd\u529b\u3002\u56e0\u6b64\uff0c\u6211\u4eec\u4ecb\u7ecdWorldDreamer\uff0c\u8fd9\u662f\u4e00\u4e2a\u5f00\u521b\u6027\u7684\u4e16\u754c\u6a21\u578b\uff0c\u65e8\u5728\u57f9\u517b\u5bf9\u4e00\u822c\u4e16\u754c\u7269\u7406\u548c\u8fd0\u52a8\u7684\u5168\u9762\u7406\u89e3\uff0c\u4ece\u800c\u663e\u8457\u589e\u5f3a\u89c6\u9891\u751f\u6210\u7684\u80fd\u529b\u3002WorldDreamer\u4ece\u5927\u578b\u8bed\u8a00\u6a21\u578b\u7684\u6210\u529f\u4e2d\u6c72\u53d6\u7075\u611f\uff0c\u5c06\u4e16\u754c\u5efa\u6a21\u5b9a\u4e49\u4e3a\u65e0\u76d1\u7763\u7684\u89c6\u89c9\u5e8f\u5217\u5efa\u6a21\u6311\u6218\u3002\u8fd9\u662f\u901a\u8fc7\u5c06\u89c6\u89c9\u8f93\u5165\u6620\u5c04\u5230\u79bb\u6563\u7684\u4ee4\u724c\u5e76\u9884\u6d4b\u63a9\u7801\u6765\u5b9e\u73b0\u7684\u3002\u5728\u6b64\u8fc7\u7a0b\u4e2d\uff0c\u6211\u4eec\u7ed3\u5408\u4e86\u591a\u6a21\u5f0f\u63d0\u793a\uff0c\u4ee5\u4fc3\u8fdb\u4e16\u754c\u6a21\u578b\u5185\u7684\u4ea4\u4e92\u3002\u6211\u4eec\u7684\u5b9e\u9a8c\u8868\u660e\uff0cWorldDreamer\u5728\u751f\u6210\u4e0d\u540c\u573a\u666f\u7684\u89c6\u9891\u65b9\u9762\u8868\u73b0\u51fa\u8272\uff0c\u5305\u62ec\u81ea\u7136\u573a\u666f\u548c\u9a7e\u9a76\u73af\u5883\u3002WorldDreamer\u5c55\u793a\u4e86\u5728\u6267\u884c\u6587\u672c\u5230\u89c6\u9891\u8f6c\u6362\u3001\u56fe\u50cf\u5230\u89c6\u9891\u5408\u6210\u548c\u89c6\u9891\u7f16\u8f91\u7b49\u4efb\u52a1\u65b9\u9762\u7684\u591a\u529f\u80fd\u6027\u3002\u8fd9\u4e9b\u7ed3\u679c\u7a81\u663e\u4e86WorldDreamer\u5728\u6355\u6349\u4e0d\u540c\u4e00\u822c\u4e16\u754c\u73af\u5883\u4e2d\u7684\u52a8\u6001\u5143\u7d20\u65b9\u9762\u7684\u6709\u6548\u6027\u3002<\/p>\n

                         <\/p>\n

                        Think2Drive: Efficient Reinforcement Learning by Thinking in Latent World Model for Quasi-Realistic Autonomous Driving (in CARLA-v2)<\/h3>\n
                          \n
                        • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2402.16720v2<\/li>\n<\/ul>\n

                          \u4e0a\u4ea4\u7684\u5de5\u4f5c\uff1a\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u81ea\u52a8\u9a7e\u9a76\uff08AD\uff09\uff0c\u5c24\u5176\u662f\u57ce\u5e02\u9a7e\u9a76\uff0c\u6d89\u53ca\u8bb8\u591a\u5f2f\u9053\u60c5\u51b5\u3002\u6700\u8fd1\u53d1\u5e03\u7684AD\u6a21\u62df\u5668CARLA v2\u5728\u9a7e\u9a76\u573a\u666f\u4e2d\u589e\u52a0\u4e8639\u4e2a\u5e38\u89c1\u4e8b\u4ef6\uff0c\u4e0eCARLA v1\u76f8\u6bd4\u63d0\u4f9b\u4e86\u66f4\u903c\u771f\u7684\u6d4b\u8bd5\u5e73\u53f0\u3002\u5b83\u7ed9\u793e\u533a\u5e26\u6765\u4e86\u65b0\u7684\u6311\u6218\uff0c\u5230\u76ee\u524d\u4e3a\u6b62\uff0c\u8fd8\u6ca1\u6709\u6587\u732e\u62a5\u9053V2\u4e2d\u7684\u65b0\u573a\u666f\u53d6\u5f97\u4e86\u4efb\u4f55\u6210\u529f\uff0c\u56e0\u4e3a\u73b0\u6709\u7684\u5de5\u4f5c\u5927\u591a\u5fc5\u987b\u4f9d\u8d56\u4e8e\u7279\u5b9a\u7684\u89c4\u5212\u89c4\u5219\uff0c\u4f46\u5b83\u4eec\u65e0\u6cd5\u6db5\u76d6CARLA V2\u4e2d\u66f4\u590d\u6742\u7684\u6848\u4f8b\u3002\u5728\u8fd9\u9879\u5de5\u4f5c\u4e2d\uff0c\u6211\u4eec\u4e3b\u52a8\u76f4\u63a5\u8bad\u7ec3\u4e00\u4e2a\u89c4\u5212\u8005\uff0c\u5e0c\u671b\u7075\u6d3b\u6709\u6548\u5730\u5904\u7406\u6781\u7aef\u60c5\u51b5\uff0c\u6211\u4eec\u8ba4\u4e3a\u8fd9\u4e5f\u662fAD\u7684\u672a\u6765\u3002\u636e\u6211\u4eec\u6240\u77e5\uff0c\u6211\u4eec\u5f00\u53d1\u4e86\u7b2c\u4e00\u4e2a\u57fa\u4e8e\u6a21\u578b\u7684RL\u65b9\u6cd5\uff0c\u540d\u4e3aThink2Drive for AD\uff0c\u4f7f\u7528\u4e16\u754c\u6a21\u578b\u6765\u5b66\u4e60\u73af\u5883\u7684\u8f6c\u53d8\uff0c\u7136\u540e\u5b83\u5145\u5f53\u795e\u7ecf\u6a21\u62df\u5668\u6765\u8bad\u7ec3\u89c4\u5212\u8005\u3002\u7531\u4e8e\u4f4e\u7ef4\u72b6\u6001\u7a7a\u95f4\u548c\u4e16\u754c\u6a21\u578b\u4e2d\u5f20\u91cf\u7684\u5e76\u884c\u8ba1\u7b97\uff0c\u8fd9\u79cd\u8303\u5f0f\u663e\u8457\u63d0\u9ad8\u4e86\u8bad\u7ec3\u6548\u7387\u3002\u56e0\u6b64\uff0cThink2Drive\u80fd\u591f\u5728\u5355\u4e2aA6000 GPU\u4e0a\u8bad\u7ec33\u5929\u5185\u4ee5\u4e13\u5bb6\u7ea7\u719f\u7ec3\u7a0b\u5ea6\u8fd0\u884cCARLA v2\uff0c\u636e\u6211\u4eec\u6240\u77e5\uff0c\u5230\u76ee\u524d\u4e3a\u6b62\uff0cCARLA v2\u4e0a\u8fd8\u6ca1\u6709\u6210\u529f\u7684\u62a5\u544a\uff08100%\u7684\u8def\u7ebf\u5b8c\u6210\uff09\u3002\u6211\u4eec\u8fd8\u63d0\u51fa\u4e86CornerCase Repository\uff0c\u8fd9\u662f\u4e00\u4e2a\u652f\u6301\u6309\u573a\u666f\u8bc4\u4f30\u9a7e\u9a76\u6a21\u578b\u7684\u57fa\u51c6\u3002\u6b64\u5916\uff0c\u6211\u4eec\u63d0\u51fa\u4e86\u4e00\u79cd\u65b0\u7684\u5e73\u8861\u6307\u6807\uff0c\u901a\u8fc7\u8def\u7ebf\u5b8c\u6210\u60c5\u51b5\u3001\u8fdd\u89c4\u6b21\u6570\u548c\u573a\u666f\u5bc6\u5ea6\u6765\u8bc4\u4f30\u6027\u80fd\uff0c\u4ee5\u4fbf\u9a7e\u9a76\u5206\u6570\u53ef\u4ee5\u63d0\u4f9b\u66f4\u591a\u5173\u4e8e\u5b9e\u9645\u9a7e\u9a76\u6027\u80fd\u7684\u4fe1\u606f\u3002<\/p>\n

                           <\/p>\n

                          OccWorld: Learning a 3D Occupancy World Model for Autonomous Driving<\/h3>\n
                            \n
                          • \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2311.16038v1<\/li>\n
                          • \u5f00\u6e90\u94fe\u63a5\uff1ahttps:\/\/github.com\/wzzheng\/OccWorld<\/li>\n<\/ul>\n

                            \u6e05\u534e\u56e2\u961f\u7684\u5de5\u4f5c\uff1a\u4e86\u89e33D\u573a\u666f\u5982\u4f55\u6f14\u53d8\u5bf9\u4e8e\u81ea\u52a8\u9a7e\u9a76\u51b3\u7b56\u81f3\u5173\u91cd\u8981\u3002\u5927\u591a\u6570\u73b0\u6709\u65b9\u6cd5\u901a\u8fc7\u9884\u6d4b\u5bf9\u8c61\u6846\u7684\u8fd0\u52a8\u6765\u5b9e\u73b0\u8fd9\u4e00\u70b9\uff0c\u8fd9\u65e0\u6cd5\u6355\u83b7\u66f4\u7ec6\u7c92\u5ea6\u7684\u573a\u666f\u4fe1\u606f\u3002\u672c\u6587\u63a2\u7d22\u4e86\u4e00\u79cd\u57283D\u5360\u7528\u7a7a\u95f4\u4e2d\u5b66\u4e60\u4e16\u754c\u6a21\u578bOccWorld\u7684\u65b0\u6846\u67b6\uff0c\u4ee5\u540c\u65f6\u9884\u6d4b\u81ea\u8f66\u7684\u8fd0\u52a8\u548c\u5468\u56f4\u573a\u666f\u7684\u6f14\u53d8\u3002\u6211\u4eec\u5efa\u8bae\u57fa\u4e8e3D\u5360\u7528\u800c\u4e0d\u662f3D\u8fb9\u754c\u6846\u548c\u5206\u5272\u56fe\u6765\u5b66\u4e60\u4e16\u754c\u6a21\u578b\uff0c\u539f\u56e0\u6709\u4e09\uff1a1\uff09\u8868\u73b0\u529b\uff1a3D\u5360\u7528\u53ef\u4ee5\u63cf\u8ff0\u573a\u666f\u7684\u66f4\u7ec6\u7c92\u5ea6\u76843D\u7ed3\u6784\uff1b2\uff09\u6548\u7387\uff1a\u83b7\u5f973D\u5360\u7528\u7387\u66f4\u7ecf\u6d4e\uff08\u4f8b\u5982\uff0c\u4ece\u7a00\u758f\u7684LiDAR\u70b9\uff09\u30023\uff09\u591a\u529f\u80fd\u6027\uff1a3D\u5360\u7528\u53ef\u4ee5\u9002\u5e94\u89c6\u89c9\u548c\u6fc0\u5149\u96f7\u8fbe\u3002\u4e3a\u4e86\u4fbf\u4e8e\u5bf9\u4e16\u754c\u6f14\u5316\u8fdb\u884c\u5efa\u6a21\uff0c\u6211\u4eec\u5b66\u4e60\u4e86\u4e00\u79cd\u57fa\u4e8e\u91cd\u5efa\u76843D\u5360\u7528\u573a\u666f\u6807\u8bb0\u5668\uff0c\u4ee5\u83b7\u5f97\u79bb\u6563\u7684\u573a\u666f\u6807\u8bb0\u6765\u63cf\u8ff0\u5468\u56f4\u7684\u573a\u666f\u3002\u7136\u540e\uff0c\u6211\u4eec\u91c7\u7528\u7c7b\u4f3cGPT\u7684\u65f6\u7a7a\u751f\u6210Transformer\u6765\u751f\u6210\u540e\u7eed\u573a\u666f\u548c\u81ea\u8f66\u4ee4\u724c\uff0c\u4ee5\u89e3\u7801\u672a\u6765\u7684\u5360\u7528\u548c\u81ea\u8f66\u8f68\u8ff9\u3002\u5728\u5e7f\u6cdb\u4f7f\u7528\u7684nuScenes\u57fa\u51c6\u4e0a\u8fdb\u884c\u7684\u5e7f\u6cdb\u5b9e\u9a8c\u8bc1\u660e\u4e86OccWorld\u6709\u6548\u6a21\u62df\u9a7e\u9a76\u573a\u666f\u6f14\u53d8\u7684\u80fd\u529b\u3002OccWorld\u8fd8\u53ef\u4ee5\u5728\u4e0d\u4f7f\u7528\u5b9e\u4f8b\u548c\u5730\u56fe\u76d1\u7763\u7684\u60c5\u51b5\u4e0b\u751f\u6210\u5177\u6709\u7ade\u4e89\u529b\u7684\u89c4\u5212\u7ed3\u679c\u3002<\/p>\n

                             <\/p>\n

                            \u4ece\u8fd9\u4e9b\u5de5\u4f5c\u4e2d\u6211\u4eec\u53ef\u4ee5\u603b\u7ed3\u51fa\u4ee5\u4e0b\u51e0\u70b9\uff1a<\/p>\n

                              \n
                            1. \u57fa\u4e8e\u4e16\u754c\u6a21\u578b\u7684\u7aef\u5230\u7aef\u65b9\u6cd5\u8fd8\u5728\u6301\u7eed\u53d1\u5c55\uff0c\u672a\u6765\u80fd\u5426\u843d\u5730\u8fd8\u9700\u89c2\u671b\uff1b<\/li>\n
                            2. Occ\u4efb\u52a1\u53ef\u4ee5\u65e0\u7f1d\u5730\u540c\u4e16\u754c\u6a21\u578b\u76f8\u7ed3\u5408\uff0c\u5f15\u5165VLM\u548c3D GS\u4e5f\u662f\u672a\u6765\u53ef\u4ee5\u8fdb\u4e00\u6b65\u6269\u5c55\u7814\u7a76\u7684\u65b9\u5411\uff1b<\/li>\n
                            3. \u4e16\u754c\u6a21\u578b\u7684\u6838\u5fc3\u662f\u751f\u6210\u3002\u57fa\u4e8e\u751f\u6210\u7684\u6846\u67b6\uff0c\u53ef\u4ee5\u6dfb\u52a0\u5f88\u591a\u6a21\u5757\uff0c\u4ee5\u8fdb\u4e00\u6b65\u589e\u5f3a\u53ef\u662f\u89e3\u91ca\u6027\u548c\u63d0\u5347\u6027\u80fd\u3002<\/li>\n
                            4. \u5f53\u4e0b\u4e16\u754c\u6a21\u578b\u5bf9\u8fd0\u52a8\u89c4\u5f8b\u3001\u7269\u7406\u89c4\u5219\u7684\u5efa\u6a21\u80fd\u529b\u8fd8\u6709\u9650\u3002<\/li>\n<\/ol>\n<\/div>\n<\/div>\n
                              <\/div>\n

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

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

                              \u5f53\u4e0b\u81ea\u52a8\u9a7e\u9a76\u65b9\u5411\u7684\u4e16\u754c\u6a21\u578b\u53ef\u4ee5\u5206\u6210\u4e24\u5927\u7c7b\uff1a\u751f\u6210\u5f0f\u548c\u7aef\u5230\u7aef\u3002 \u672c\u6587\u7ecf\u81ea\u52a8\u9a7e\u9a76\u4e4b\u5fc3\u516c\u4f17\u53f7\u6388\u6743\u8f6c\u8f7d\uff0c\u8f6c\u8f7d\u8bf7\u8054\u7cfb\u51fa\u5904\u3002 […]<\/p>\n

                              <\/div>\n

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

                              <\/div>\n","protected":false},"author":56,"featured_media":22631,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[23,20,80],"tags":[1374],"class_list":["post-22629","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-23","category-20","category-80","tag-1374"],"_links":{"self":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/22629","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=22629"}],"version-history":[{"count":1,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/22629\/revisions"}],"predecessor-version":[{"id":22632,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/22629\/revisions\/22632"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media\/22631"}],"wp:attachment":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media?parent=22629"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/categories?post=22629"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/tags?post=22629"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}