{"id":507,"date":"2025-10-11T01:47:06","date_gmt":"2025-10-10T17:47:06","guid":{"rendered":"https:\/\/www.ndnlab.com\/?p=507"},"modified":"2025-11-09T04:36:32","modified_gmt":"2025-11-08T20:36:32","slug":"icnp-2025%e8%ae%ba%e6%96%87%e6%8a%a5%e9%81%93-%e4%b9%8b-ai4net","status":"publish","type":"post","link":"https:\/\/www.ndnlab.com\/?p=507","title":{"rendered":"ICNP 2025\u8bba\u6587\u62a5\u9053 \u4e4b AI4Net"},"content":{"rendered":"\n

\u7b2c 33 \u5c4a IEEE ICNP \u4e8e 2025 \u5e74 9 \u6708\u5728\u97e9\u56fd\u9996\u5c14\u53ec\u5f00\uff0c\u672c\u6b21\u5927\u4f1a\u5171\u8ba1 238 \u7bc7\u6295\u7a3f\uff0c\u5f55\u7528 60 \u7bc7\uff08\u5f55\u7528\u7387\u4e3a 25%\uff09\u3002<\/p>\n\n\n\n

\u300aMamba4Net: \u84b8\u998f\u5f0f\u6df7\u5408Mamba\u5927\u8bed\u8a00\u6a21\u578b\u5728\u7f51\u7edc\u9886\u57df\u7684\u5e94\u7528\u300b\u8bba\u6587\u4ecb\u7ecd<\/h3>\n\n\n\n

\u300aMamba4Net: Distilled Hybrid Mamba Large Language Models For Networking\u300b\u662f\u4e00\u7bc7\u53d1\u8868\u4e8e2025\u5e74IEEE\u7b2c33\u5c4a\u56fd\u9645\u7f51\u7edc\u534f\u8bae\u4f1a\u8bae\uff08ICNP\uff09\u7684\u8bba\u6587<\/a>\uff0c\u5317\u4eac\u5927\u5b66\u6df1\u5733\u7814\u7a76\u751f\u9662ICNLab\u3002\u8be5\u8bba\u6587\u805a\u7126\u4e8e\u5927\u8bed\u8a00\u6a21\u578b\uff08LLMs\uff09\u5728\u7f51\u7edc\u9886\u57df\u7684\u5e94\u7528\uff0c\u9488\u5bf9Transformer-based LLMs\u7684\u8ba1\u7b97\u5f00\u9500\u548c\u5185\u5b58\u6d88\u8017\u95ee\u9898\uff0c\u63d0\u51fa\u4e86\u4e00\u79cd\u521b\u65b0\u7684\u8de8\u67b6\u6784\u77e5\u8bc6\u84b8\u998f\u6846\u67b6Mamba4Net\u3002\u8be5\u6846\u67b6\u901a\u8fc7\u5c06Transformer LLMs\u7684\u7f51\u7edc\u7279\u5b9a\u77e5\u8bc6\u8f6c\u79fb\u5230\u7ebf\u6027\u65f6\u95f4\u590d\u6742\u5ea6\u7684Mamba\u67b6\u6784\u5b66\u751f\u6a21\u578b\u4e2d\uff0c\u5b9e\u73b0\u9ad8\u6548\u7684\u6a21\u578b\u538b\u7f29\u548c\u6027\u80fd\u63d0\u5347\u3002\u8bba\u6587\u5f3a\u8c03\u4e86\u5728\u8d44\u6e90\u53d7\u9650\u73af\u5883\u4e0b\u7684\u5b9e\u9645\u90e8\u7f72\u6f5c\u529b\uff0c\u5e76\u5f00\u6e90\u4e86\u6e90\u4ee3\u7801\uff0c\u4ee5\u4fc3\u8fdb\u8fdb\u4e00\u6b65\u7814\u7a76\u3002<\/p>\n\n\n\n

\"ICNP<\/figure>\n\n\n\n

\u8bba\u6587\u7684\u80cc\u666f\u6e90\u4e8e\u5168\u7403\u6570\u636e\u6d41\u91cf\u7684\u6307\u6570\u7ea7\u589e\u957f\uff0c\u8fd9\u8981\u6c42\u7f51\u7edc\u7cfb\u7edf\u4f18\u5316\u4ee5\u786e\u4fdd\u65e0\u7f1d\u8fde\u63a5\u548c\u6027\u80fd\u3002\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u5df2\u88ab\u5e7f\u6cdb\u7528\u4e8e\u5e26\u5bbd\u9884\u6d4b\u3001\u62e5\u585e\u63a7\u5236\u548c\u6d41\u91cf\u5206\u6790\u7b49\u4efb\u52a1\uff0c\u4f46\u4f20\u7edf\u6a21\u578b\u9002\u5e94\u4e0d\u540c\u4efb\u52a1\u65f6\u5de5\u7a0b\u6210\u672c\u9ad8\uff0c\u4e14\u5728\u65b0\u6570\u636e\u5206\u5e03\u4e0b\u7684\u9c81\u68d2\u6027\u5dee\u3002\u4e3a\u89e3\u51b3\u8fd9\u4e9b\u95ee\u9898\uff0c\u8fd1\u5e74\u6765\u7814\u7a76\u8005\u63a2\u7d22\u4e86Transformer-based LLMs\u5728\u7f51\u7edc\u4e2d\u7684\u5e94\u7528\uff0c\u79f0\u4e3aLLMs for Networks (LLMs4Net)\u3002\u8fd9\u4e9b\u65b9\u6cd5\u5229\u7528LLMs\u7684\u9884\u8bad\u7ec3\u77e5\u8bc6\uff0c\u4ee5\u5355\u4e00\u6a21\u578b\u89e3\u51b3\u591a\u4e2a\u7f51\u7edc\u95ee\u9898\uff0c\u663e\u8457\u964d\u4f4e\u5de5\u7a0b\u6210\u672c\uff0c\u5e76\u5c55\u73b0\u51fa\u826f\u597d\u6027\u80fd\u3002\u7136\u800c\uff0c\u73b0\u6709\u7684LLMs4Net\u65b9\u6cd5\u5b58\u5728\u4e24\u5927\u5c40\u9650\uff1a\u4e00\u662fTransformer\u7684\u4e8c\u6b21\u65f6\u95f4\u590d\u6742\u5ea6\u5bfc\u81f4\u8ba1\u7b97\u5f00\u9500\u5de8\u5927\uff0c\u4e0d\u9002\u5408\u5904\u7406\u5927\u89c4\u6a21\u8fde\u7eed\u7f51\u7edc\u6570\u636e\u6d41\uff1b\u4e8c\u662f\u6a21\u578b\u53c2\u6570\u5e9e\u5927\uff0c\u5305\u542b\u5927\u91cf\u975e\u7f51\u7edc\u76f8\u5173\u77e5\u8bc6\uff0c\u5bfc\u81f4\u5185\u5b58\u5360\u7528\u8fc7\u9ad8\uff0c\u96be\u4ee5\u90e8\u7f72\u5728\u8def\u7531\u5668\u6216AR\u5934\u663e\u7b49\u8d44\u6e90\u53d7\u9650\u8bbe\u5907\u4e0a\u3002<\/p>\n\n\n\n

\u53d7Deepseek-R1\u6a21\u578b\u5728\u77e5\u8bc6\u84b8\u998f\u8303\u5f0f\u4e0b\u7684\u9ad8\u6548\u6027\u548c\u6027\u80fd\u542f\u53d1\uff0c\u8bba\u6587\u5f15\u5165Mamba4Net\u6846\u67b6\u3002\u8be5\u6846\u67b6\u7684\u6838\u5fc3\u662f\u8de8\u67b6\u6784\u84b8\u998f\uff0c\u5c06Transformer LLMs\u4f5c\u4e3a\u6559\u5e08\u6a21\u578b\uff0c\u63d0\u53d6\u7f51\u7edc\u7279\u5b9a\u77e5\u8bc6\uff0c\u5e76\u8f6c\u79fb\u5230\u57fa\u4e8eMamba\u7684\u5b66\u751f\u6a21\u578b\u4e2d\u3002Mamba\u67b6\u6784\u901a\u8fc7\u9009\u62e9\u6027\u72b6\u6001\u7a7a\u95f4\u6a21\u578b\uff08SSMs\uff09\u5b9e\u73b0\u7ebf\u6027\u65f6\u95f4\u590d\u6742\u5ea6\uff0c\u80fd\u591f\u6355\u6349\u5168\u5c40\u4e0a\u4e0b\u6587\uff0c\u540c\u65f6\u663e\u8457\u964d\u4f4e\u8ba1\u7b97\u9700\u6c42\u3002\u5b66\u751f\u6a21\u578b\u91c7\u7528\u6df7\u5408\u8bbe\u8ba1\uff0c\u4e3b\u8981\u7531Mamba\u5757\u7ec4\u6210\uff0c\u5e76\u5728\u8f93\u51fa\u5c42\u9644\u8fd1\u4fdd\u7559\u5c11\u91cfTransformer\u5757\uff0c\u4ee5\u7ef4\u6301\u84b8\u998f\u8fc7\u7a0b\u7684\u7a33\u5b9a\u6027\u3002\u76f8\u6bd4\u6559\u5e08\u6a21\u578b\uff0c\u5b66\u751f\u6a21\u578b\u5c42\u6570\u66f4\u5c11\u3001\u4ee4\u724c\u8868\u793a\u7ef4\u5ea6\u66f4\u5c0f\uff0c\u4ece\u800c\u5b9e\u73b0\u6a21\u578b\u89c4\u6a21\u538b\u7f29\u548c\u5185\u5b58\u4f18\u5316\u3002<\/p>\n\n\n\n

\u8bba\u6587\u7684\u4e3b\u8981\u8d21\u732e\u5305\u62ec\u4e09\u4e2a\u65b9\u9762\u3002\u9996\u5148\uff0c\u63d0\u51fa\u9886\u57df\u77e5\u8bc6\u5bfc\u5411\u7684\u8de8\u5f02\u6784\u84b8\u998f\uff08Domain Knowledge-Oriented Cross-Heterogeneous Distillation, DKO\uff09\u3002DKO\u9488\u5bf9\u76d1\u7763\u5b66\u4e60\uff08SL\uff09\u548c\u5f3a\u5316\u5b66\u4e60\uff08RL\uff09\u4e24\u79cd\u7f51\u7edc\u8303\u5f0f\u8bbe\u8ba1\u84b8\u998f\u673a\u5236\u3002\u5728SL\u4e2d\uff08\u5982\u89c6\u53e3\u9884\u6d4b\u4efb\u52a1\uff09\uff0c\u5b66\u751f\u6a21\u578b\u7684\u635f\u5931\u51fd\u6570\u7ed3\u5408\u76d1\u7763\u635f\u5931\u548c\u6559\u5e08-\u5b66\u751f\u8f93\u51fa\u5206\u5e03\u7684KL\u6563\u5ea6\uff1b\u5728RL\u4e2d\uff08\u5982\u81ea\u9002\u5e94\u6bd4\u7279\u7387\u6d41\u5a92\u4f53\u548c\u96c6\u7fa4\u4f5c\u4e1a\u8c03\u5ea6\uff09\uff0c\u6dfb\u52a0\u7b56\u7565\u84b8\u998fKL\u9879\uff0c\u4ee5\u5339\u914d\u6559\u5e08\u7b56\u7565\u3002\u540c\u65f6\uff0c\u4f7f\u7528\u4f4e\u79e9\u9002\u5e94\uff08LoRA\uff09\u9ad8\u6548\u5fae\u8c03\u6559\u5e08\u6a21\u578b\uff0c\u4ec5\u66f4\u65b0\u5c11\u91cf\u53c2\u6570\uff0c\u5c06\u6743\u91cd\u77e9\u9635\u5206\u89e3\u4e3a\u4f4e\u79e9\u77e9\u9635\u4e58\u79ef\uff0c\u51cf\u5c11\u8ba1\u7b97\u5f00\u9500\u3002\u5176\u6b21\uff0c\u5f15\u5165\u8de8\u5f02\u6784\u6743\u91cd\u91cd\u7528\uff08Cross-Heterogeneous Weight Reusing, CWR\uff09\u3002\u4e0d\u540c\u4e8e\u968f\u673a\u521d\u59cb\u5316\u5b66\u751f\u6a21\u578b\uff0cCWR\u901a\u8fc7\u5206\u89e3\u6559\u5e08\u6ce8\u610f\u529b\u5757\u7684\u9ad8\u7ef4\u77e9\u9635\u4e3a\u4f4e\u79e9\u5b50\u7a7a\u95f4\uff0c\u751f\u6210\u7d27\u51d1\u56e0\u5b50\uff0c\u7528\u4e8e\u521d\u59cb\u5316Mamba\u5b66\u751f\u53c2\u6570\u3002\u8fd9\u7f29\u5c0f\u4e86\u8868\u793a\u5dee\u8ddd\uff0c\u63d0\u9ad8\u84b8\u998f\u7a33\u5b9a\u6027\u548c\u6548\u7387\uff0c\u907f\u514d\u4e0d\u5fc5\u8981\u5f00\u9500\u3002\u7b2c\u4e09\uff0c\u8fdb\u884c\u5168\u9762\u5b9e\u9a8c\u5e76\u5f00\u6e90\u4ee3\u7801\uff0c\u5728\u89c6\u53e3\u9884\u6d4b\uff08VP\uff09\u3001\u81ea\u9002\u5e94\u6bd4\u7279\u7387\u6d41\u5a92\u4f53\uff08ABR\uff09\u548c\u96c6\u7fa4\u4f5c\u4e1a\u8c03\u5ea6\uff08CJS\uff09\u4e09\u4e2a\u4efb\u52a1\u4e0a\u9a8c\u8bc1Mamba4Net\u3002<\/p>\n\n\n\n

\u65b9\u6cd5\u8bba\u90e8\u5206\u8be6\u7ec6\u9610\u8ff0\u4e86Mamba4Net\u7684\u6574\u4f53\u7ba1\u9053\uff0c\u5305\u62ec\u591a\u6a21\u6001\u7f16\u7801\u5668\u3001\u4efb\u52a1\u7279\u5b9a\u7f51\u7edc\u5934\u548c\u6559\u5e08-\u5b66\u751f\u6846\u67b6\u3002\u591a\u6a21\u6001\u7f16\u7801\u5668\u5904\u7406\u89c6\u9891\u3001\u56fe\u50cf\u6216\u56fe\u6570\u636e\u7b49\u5f02\u6784\u8f93\u5165\uff0c\u4f7f\u7528\u9884\u8bad\u7ec3\u7f16\u7801\u5668\uff08\u5982Visual Transformer for\u56fe\u50cf\uff0cGraph Neural Networks for\u56fe\u6570\u636e\uff09\u8f6c\u6362\u4e3a\u7edf\u4e00\u4ee4\u724c\u5d4c\u5165\uff0c\u5e76\u901a\u8fc7\u7ebf\u6027\u5c42\u6295\u5f71\u5230LLMs\u6216Mamba\u7684\u7ef4\u5ea6\u7a7a\u95f4\u3002\u4efb\u52a1\u7279\u5b9a\u5934\u66ff\u6362\u9ed8\u8ba4LM\u5934\uff0c\u76f4\u63a5\u6295\u5f71\u8f93\u51fa\u4e3a\u6807\u51c6\u5316\u683c\u5f0f\uff0c\u7f13\u89e3\u6a21\u578b\u5e7b\u89c9\u95ee\u9898\u3002\u6559\u5e08\u6a21\u578b\u4e3a\u9884\u8bad\u7ec3Transformer LLM\uff0c\u5b66\u751f\u4e3a\u6df7\u5408Mamba\u6a21\u578b\uff0c\u901a\u8fc7DKO\u548cCWR\u5b9e\u73b0\u77e5\u8bc6\u8f6c\u79fb\u3002<\/p>\n\n\n\n

\u5b9e\u9a8c\u5728\u4e09\u4e2a\u4ee3\u8868\u6027\u4efb\u52a1\u4e0a\u8bc4\u4f30Mamba4Net\u3002VP\u4efb\u52a1\u4f7f\u7528360\u00b0\u89c6\u9891\u6570\u636e\u96c6\uff0c\u9884\u6d4b\u7528\u6237\u89c6\u53e3\u4ee5\u4f18\u5316\u5e26\u5bbd\uff1bABR\u4efb\u52a1\u6a21\u62df\u6d41\u5a92\u4f53\u73af\u5883\uff0c\u9009\u62e9\u6bd4\u7279\u7387\u4ee5\u6700\u5927\u5316QoE\uff1bCJS\u4efb\u52a1\u5904\u7406\u96c6\u7fa4\u8c03\u5ea6\uff0c\u4f18\u5316\u4f5c\u4e1a\u5b8c\u6210\u65f6\u95f4\u3002\u7ed3\u679c\u663e\u793a\uff0cMamba4Net\u5728\u6027\u80fd\u4e0a\u4f18\u4e8e\u4e0d\u4f7f\u7528LLMs\u7684\u73b0\u6709\u65b9\u6cd5\uff0c\u5982\u5728VP\u4e2d\u51c6\u786e\u7387\u66f4\u9ad8\uff0c\u5728ABR\u4e2dQoE\u63d0\u5347\uff0c\u5728CJS\u4e2d\u4f5c\u4e1a\u5b8c\u6210\u65f6\u95f4\u7f29\u77ed\u3002\u76f8\u6bd4\u76f4\u63a5\u5e94\u7528Transformer LLMs\uff0cMamba4Net\u5b9e\u73b0\u663e\u8457\u6548\u7387\u63d0\u5347\uff1a\u541e\u5410\u91cf\u63d0\u9ad83.96\u500d\uff0c\u5b58\u50a8\u5360\u7528\u4ec5\u4e3a5.48%\u3002\u5728\u8d44\u6e90\u53d7\u9650\u573a\u666f\u4e0b\uff0c\u4f18\u52bf\u66f4\u660e\u663e\uff0c\u8bc1\u660e\u4e86\u5176\u5728\u8fb9\u7f18\u8bbe\u5907\u4e0a\u7684\u6f5c\u529b\u3002\u8bba\u6587\u8fd8\u6bd4\u8f83\u4e86\u4e0d\u540c\u84b8\u998f\u7b56\u7565\uff0c\u9a8c\u8bc1\u4e86CWR\u5bf9\u8bad\u7ec3\u7a33\u5b9a\u6027\u7684\u8d21\u732e\u3002<\/p>\n\n\n\n

\u8ba8\u8bba\u90e8\u5206\u5206\u6790\u4e86Mamba4Net\u7684\u5c40\u9650\uff0c\u5982\u84b8\u998f\u8fc7\u7a0b\u53ef\u80fd\u4e22\u5931\u90e8\u5206\u590d\u6742\u77e5\u8bc6\uff0c\u4ee5\u53ca\u672a\u6765\u6269\u5c55\u5230\u66f4\u591a\u7f51\u7edc\u4efb\u52a1\u7684\u6f5c\u529b\u3002\u7ed3\u8bba\u5f3a\u8c03\uff0c\u8be5\u6846\u67b6\u4e3aLLMs\u5728\u7f51\u7edc\u9886\u57df\u7684\u6210\u672c\u6709\u6548\u5e94\u7528\u5f00\u8f9f\u65b0\u8def\u5f84\uff0c\u901a\u8fc7\u7ebf\u6027\u590d\u6742\u5ea6\u548c\u6a21\u578b\u538b\u7f29\uff0c\u5b9e\u73b0\u9ad8\u6027\u80fd\u4e0e\u9ad8\u6548\u7684\u5e73\u8861\u3002\u8be5\u8bba\u6587\u4e0d\u4ec5\u63a8\u8fdb\u4e86LLMs4Net\u7684\u7814\u7a76\uff0c\u8fd8\u4e3a\u8de8\u67b6\u6784\u77e5\u8bc6\u84b8\u998f\u63d0\u4f9b\u65b0\u8303\u5f0f\uff0c\u5177\u6709\u91cd\u8981\u7684\u5b66\u672f\u548c\u5b9e\u7528\u4ef7\u503c\u3002<\/p>\n\n\n\n

\u7531 DANCE GROUP \u90d1\u9648\u7ca4\u5e26\u6765\u4f1a\u8bae\u62a5\u9053\u3002<\/p>\n\n\n\n

Copo: Joint Cost and Performance Optimization for Task Placement in Geo-Distributed Clouds<\/p>\n\n\n\n

Bichen Wang, Jingzhou Wang, Yu-E Sun, He Huang<\/p>\n\n\n\n

  Soochow University, China <\/p>\n\n\n\n

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

\u4ecb\u7ecd<\/strong><\/p>\n\n\n\n

\u4e91\u8ba1\u7b97\u5feb\u901f\u53d1\u5c55\u4e0b\uff0c\u5934\u90e8\u4e91\u670d\u52a1\u5382\u5546\u901a\u5e38\u5efa\u8bbe\u5730\u7406\u5206\u5e03\u5f0f\u7684\u6570\u636e\u4e2d\u5fc3\uff0c\u4ee5 \u201c\u533a\u57df\u201d \u67b6\u6784\u6765\u90e8\u7f72\u96c6\u7fa4\u652f\u6491\u5168\u7403\u670d\u52a1\u3002\u4f46\u8fd9\u6837\u505a\u7684\u670d\u52a1\u6210\u672c\u9ad8\u6602\uff1a\u4e00\u662f\u8de8\u5730\u57df\u4efb\u52a1\u90e8\u7f72\u7684\u8d44\u6e90\u6d88\u8017\uff0c\u4e8c\u662f\u8de8\u533a\u57df\u901a\u4fe1\u7684\u5e26\u5bbd\u652f\u51fa\u3002\u800c\u73b0\u6709\u65b9\u6848\u5b58\u5728\u77ed\u677f\uff0c\u8981\u4e48\u4ec5\u4f18\u5316\u5355\u4e00\u6210\u672c\uff0c\u8981\u4e48\u5ffd\u89c6\u6027\u80fd\u6307\u6807\u3002\u56e0\u6b64\uff0c\u672c\u6587\u63d0\u51faCopo \u6846\u67b6\uff0c\u901a\u8fc7\u4efb\u52a1\u90e8\u7f72\u4e0e\u6d41\u91cf\u4f20\u8f93\u8054\u5408\u51b3\u7b56\uff0c\u5b9e\u73b0\u5730\u7406\u5206\u5e03\u5f0f\u4e91\u7684\u6210\u672c\u4e0e\u6027\u80fd\u534f\u540c\u4f18\u5316\u3002<\/p>\n\n\n\n

\u6838\u5fc3\u601d\u60f3\u4e0e\u8d21\u732e<\/strong><\/p>\n\n\n\n

Copo \u7684\u6838\u5fc3\u601d\u60f3\u662f\u901a\u8fc7\u5c06\u53cc\u5c42\u4f18\u5316\u95ee\u9898\uff08\u4e0a\u5c42\u6210\u672c\u4f18\u5316\u3001\u4e0b\u5c42\u6027\u80fd\u4f18\u5316\uff09\u8f6c\u5316\u4e3a\u5355\u5c42\u4f18\u5316\u95ee\u9898\uff0c\u5e76\u8bbe\u8ba1\u4e86\u9ad8\u6548\u7684\u8fd1\u4f3c\u7b97\u6cd5\uff0c\u4ece\u800c\u5f97\u5230\u4efb\u52a1\u90e8\u7f72\u4e0e\u6d41\u91cf\u4f20\u8f93\u7684\u8054\u5408\u51b3\u7b56\u3002\u5176\u4e3b\u8981\u8d21\u732e\u4e3a\uff1a1\uff09\u901a\u8fc7\u5f15\u5165\u865a\u62df\u8282\u70b9\uff08\u4efb\u52a1\uff09\u4e0e\u865a\u62df\u8fb9\uff08\u4efb\u52a1 – \u533a\u57df\u8fde\u63a5\uff09\uff0c\u5c06\u539f\u95ee\u9898\u8f6c\u5316\u4e3a\u7ecf\u5178\u591a\u5546\u54c1\u6d41\u95ee\u9898\u30022\uff09\u5c06\u6210\u672c\u4f18\u5316\u4e0e\u6027\u80fd\u4f18\u5316\u6784\u5efa\u4e3a\u53cc\u5c42\u4f18\u5316\u6a21\u578b\uff0c\u57fa\u4e8eKKT\u6761\u4ef6\uff0c\u5c06\u4e0b\u5c42\u6027\u80fd\u4f18\u5316\u7b49\u6548\u8f6c\u5316\u4e3a\u4e0a\u5c42\u4f18\u5316\u7684\u7ea6\u675f\u6761\u4ef6\u30023\uff09\u57fa\u4e8e McCormick \u5305\u7edc\u677e\u5f1b\u65b9\u6cd5\uff0c\u5c06\u4e8c\u6b21\u89c4\u5212\u95ee\u9898\u8f6c\u5316\u4e3a\u7ebf\u6027\u89c4\u5212\u95ee\u9898\uff0c\u518d\u7ed3\u5408\u968f\u673a\u820d\u5165\u7b97\u6cd5\u6765\u6c42\u89e3\u4efb\u52a1\u90e8\u7f72\u4e0e\u6d41\u91cf\u4f20\u8f93\u7684\u8054\u5408\u51b3\u7b56\u3002<\/p>\n\n\n\n

\u5b9e\u9a8c\u8bc4\u4f30<\/strong><\/p>\n\n\n\n

\u533a\u57df\u6570\u636e\u96c6\u91c7\u7528\u534e\u4e3a\u4e91\u6570\u636e\uff0c\u4efb\u52a1\u6570\u636e\u96c6\u91c7\u7528\u963f\u91cc\u7684trace\u6570\u636e\uff0c\u8bc4\u4f30\u7535\u529b\u6210\u672c\u3001\u5e26\u5bbd\u6210\u672c\u3001\u603b\u6210\u672c\u548c\u8d1f\u8f7d\u5747\u8861\u7387\u3002\u7ed3\u679c\u8868\u660e\uff0cCopo\u7684\u7535\u529b\u6210\u672c\u548c\u5e26\u5bbd\u6210\u672c\u5747\u63a5\u8fd1\u4e8e\u5404\u81ea\u7684SOTA\u65b9\u6cd5\uff0c\u603b\u6210\u672c\u8868\u73b0\u6700\u4f18\uff0c\u8d1f\u8f7d\u5747\u8861\u63a5\u8fd1\u6700\u4f18\u3002<\/p>\n\n\n\n

\u95ee\u7b54\uff1a<\/strong><\/p>\n\n\n\n

    \n
  1. \u7528KKT\u65b9\u6cd5\u6765\u505a\uff0c\u4f18\u52bf\u5728\u54ea\u91cc\uff1f<\/li>\n<\/ol>\n\n\n\n

    A: KKT\u7684\u6838\u5fc3\u4f18\u52bf\u5728\u4e8e\u4e0a\u4e0b\u5c42\u4f18\u5316\u7edf\u4e00\u5230\u540c\u4e00\u4e2a\u6846\u67b6\u4e2d\uff0c\u4ece\u800c\u80fd\u591f\u505a\u540e\u7eed\u7684\u6c42\u89e3\u3002<\/p>\n\n\n\n

    \u4e2a\u4eba\u8bc4\u4ef7\uff1a<\/strong><\/p>\n\n\n\n

    \u672c\u6587\u6700\u7a81\u51fa\u7684\u4ef7\u503c\u5728\u4e8e\u6253\u7834\u4e86\u73b0\u6709\u4e91\u4efb\u52a1\u90e8\u7f72\u65b9\u6848 \u201c\u6210\u672c\u4f18\u5316\u201d \u4e0e \u201c\u6027\u80fd\u4f18\u5316\u201d \u5272\u88c2\u7684\u5c40\u9762\uff0c\u6784\u5efa\u4e86\u517c\u987e\u4e8c\u8005\u7684\u7edf\u4e00\u6846\u67b6\uff0c\u5e76\u63d0\u4f9b\u4e86\u5408\u9002\u7684\u6c42\u89e3\u65b9\u6cd5\u3002\u503c\u5f97\u63a2\u7a76\u7684\u662f\u8be5\u65b9\u6cd5\u662f\u5426\u80fd\u8fdb\u4e00\u6b65\u6269\u5c55\u5230\u5176\u4ed6\u9886\u57df\u3002<\/p>\n\n\n\n

    ReWeave: Traffic Engineering with Robust Path Weaving for Localized Link Failure Recovery<\/p>\n\n\n\n

    Jingyi Guan , Kun Qiu, Jin Zhao<\/p>\n\n\n\n

    Fudan University Shanghai, China<\/p>\n\n\n\n

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

    \u4ecb\u7ecd<\/strong><\/p>\n\n\n\n

    ISP \u7f51\u7edc\u4e2d\u5e38\u53d1\u751f\u94fe\u8def\u6545\u969c\uff0c\u4e3a\u6d41\u91cf\u5de5\u7a0b\uff08TE\uff09\u5f15\u5165\u6311\u6218\u3002\u73b0\u6709\u7684TE\u65b9\u6848\u6216\u8005\u4f9d\u8d56\u9759\u6001\u8def\u5f84\u91cd\u8def\u7531\uff0c\u5bfc\u81f4\u6027\u80fd\u4e0b\u964d\uff0c\u6216\u8005\u9884\u8ba1\u7b97\u51fa\u5927\u91cf\u5907\u4efd\u8def\u7531\uff0c\u4e3a\u57fa\u4e8eML\u7684\u63a8\u7406\u65b9\u6848\u5f15\u5165\u9ad8\u989d\u5f00\u9500\u3002\u56e0\u6b64\uff0c\u672c\u6587\u63d0\u51faReWeave\uff0c\u4e3a\u6bcf\u6761\u94fe\u8def\u914d\u7f6e\u4e86\u7d27\u51d1\u4e14\u4ec5\u76f8\u90bb\u8282\u70b9\u7684\u5907\u4efd\u8def\u5f84\u96c6\uff1b\u6545\u969c\u65f6\u4ec5\u6545\u969c\u94fe\u8def\u4e24\u7aef\u7684\u8def\u7531\u5668\u501f SRv6 \u7ed5\u9053\u8def\u5f84\u6765\u52a8\u6001\u91cd\u8def\u7531\u3002<\/p>\n\n\n\n

    \u6838\u5fc3\u601d\u60f3\u4e0e\u8d21\u732e<\/strong><\/p>\n\n\n\n

    ReWeave \u7684\u6838\u5fc3\u601d\u60f3\u662f\u5229\u7528\u76f8\u90bb\u8282\u70b9\u95f4\u7684\u8def\u5f84\u96c6\uff0c\u4e3a\u6bcf\u6761\u94fe\u8def\u914d\u7f6e\u672c\u5730\u5907\u4efd\u8def\u5f84\u3002\u5728\u6545\u969c\u53d1\u751f\u65f6\u523b\uff0c\u901a\u8fc7\u5728\u672c\u5730\u4fee\u590d\u539f\u59cb\u5168\u5c40\u8def\u5f84\uff0c\u4ece\u800c\u7ef4\u6301\u4e00\u4e2a\u7a33\u5065\u8def\u5f84\u96c6\uff0c\u4f7f\u9884\u8bad\u7ec3\u7684ML\u6a21\u578b\u6301\u7eed\u6709\u6548\u3002\u5177\u4f53\u800c\u8a00\uff0c\u5f53\u8def\u7531\u5668\u68c0\u6d4b\u5230\u94fe\u8def\u6545\u969c\u65f6\uff0c\u53ef\u4ee5\u501f\u52a9SDN\u901a\u8fc7\u6570\u636e\u5305\u9690\u5f0f\u63a2\u6d4b\u6765\u83b7\u53d6\u53e6\u4e00\u7aef\u8def\u7531\u5668\u7684\u5907\u7528\u8def\u5f84\uff0c\u5e76\u4fee\u6539\u6570\u636e\u5305\u7684\u6bb5\u6807\u8bc6\u7b26\uff08SID\uff09\u5217\u8868\u5b9e\u73b0\u7ed5\u9053\u8def\u7531\uff0c\u5927\u5e45\u964d\u4f4e\u4e86\u91cd\u8def\u7531\u5f00\u9500\u3002<\/p>\n\n\n\n

    \u5b9e\u9a8c\u8bc4\u4f30<\/strong><\/p>\n\n\n\n

    \u91c7\u7528\u4e94\u5c42\u5168\u8fde\u63a5\u795e\u7ecf\u7f51\u7edc\uff0c\u8bc4\u4f30Viatel\u3001Cogentco\u4e24\u4e2a\u771f\u5b9e\u7f51\u7edc\u62d3\u6251\uff0c\u4ee5\u53canorth america\u3001africa\u4e24\u4e2a\u5408\u6210\u62d3\u6251\u3002\u5728\u5355\/\u591a\u94fe\u8def\u6545\u969c\u573a\u666f\u4e0b\uff0cReWeave \u5728\u6240\u6709\u7f51\u7edc\u62d3\u6251\u4e0a\u7684\u5f52\u4e00\u5316\u94fe\u8def\u5229\u7528\u7387\u8868\u73b0\u6700\u4f73\uff0c\u4f18\u4e8e\u6240\u6709baseline\u3002\u5728\u5f00\u9500\u5206\u6790\u4e0a\uff0cReWeave\u67b6\u6784\u7b80\u5355\uff0c\u8bad\u7ec3\u65f6\u95f4\u5904\u4e8e\u524d\u5217\uff0c\u8ba1\u7b97\u65f6\u95f4\u5219\u4f18\u4e8e\u51e0\u4e4e\u6240\u6709baseline\u3002<\/p>\n\n\n\n

    \u95ee\u7b54\uff1a<\/strong><\/p>\n\n\n\n

      \n
    1. \u8fd9\u7bc7\u6587\u7ae0\u5b9e\u9a8c\u4e2d\u5e76\u6ca1\u6709\u5bf9\u6bd4LP\uff0c\u4e3a\u4ec0\u4e48\u8bf4\u6587\u7ae0\u7684\u6548\u679c\u6700\u597d\uff1f<\/li>\n<\/ol>\n\n\n\n

      A: LP\u4f5c\u4e3a\u7406\u60f3\u6700\u4f18\u89e3\uff0c\u5728\u7ed3\u679c\u4e2d\u5df2\u7ecf\u505a\u8fc7\u4e86\u5f52\u4e00\u5316\u5904\u7406\u3002\u53ef\u4ee5\u4f53\u73b0\u51fa\u672c\u65b9\u6848\u7684\u6548\u679c\u3002<\/p>\n\n\n\n

      \u4e2a\u4eba\u8bc4\u4ef7\uff1a<\/strong><\/p>\n\n\n\n

      \u6545\u969c\u6062\u590d\u7684TE\u662f\u4e00\u4e2a\u7ecf\u5178\u7684\u9886\u57df\uff0c\u8fd9\u7bc7\u5de5\u4f5c\u901a\u8fc7\u805a\u7126\u94fe\u8def\u7ea7\u7684\u672c\u5730\u5316\u6062\u590d\uff0c\u5de7\u5999\u907f\u5f00\u4e86\u4f20\u7edf\u65b9\u6cd5\u7684\u75db\u70b9\uff0c\u65e2\u65e0\u9700\u63a7\u5236\u5668\u7684\u7e41\u7410\u5e72\u9884\uff0c\u53c8\u80fd\u7ef4\u6301\u9884\u8bad\u7ec3 ML \u6a21\u578b\u7684\u6709\u6548\u6027\uff0c\u907f\u514d\u4e86\u6a21\u578b\u91cd\u65b0\u8bad\u7ec3\u7684\u9ad8\u6602\u6210\u672c\u3002\u5728\u6548\u679c\u6709\u6240\u63d0\u5347\u7684\u540c\u65f6\uff0c\u6548\u7387\u4f18\u4e8e\u5176\u4f59\u57fa\u4e8eDNN\u7684\u65b9\u6cd5\u3002<\/p>\n\n\n\n

      Planner: A Generative Graph Learning Framework for Noisy and Dynamic In-band Network Telemetry<\/p>\n\n\n\n

      Xiaoquan Zhang1, WaiMing Lau1, Lin Cui1, Fung Po Tso2, Zhuoqian Liang1, Zhen Zhang1 and Yuhui Deng1 <\/p>\n\n\n\n

       1 Jinan University, Guangzhou. <\/p>\n\n\n\n

       2 Loughborough University, UK<\/p>\n\n\n\n

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

      \u4ecb\u7ecd<\/strong><\/p>\n\n\n\n

      \u5e26\u5185\u7f51\u7edc\u9065\u6d4b\uff08INT\uff09\u901a\u8fc7\u5c06\u6570\u636e\u5d4c\u5165\u5230\u6570\u636e\u5305\u6765\u5b9e\u73b0\u5b9e\u65f6\u76d1\u63a7\uff0c\u53ef\u7f16\u7a0b\u4ea4\u6362\u673a\u8fdb\u4e00\u6b65\u63d0\u5347\u4e86INT\u5728\u786c\u4ef6\u4e0a\u52a8\u6001\u5b9a\u5236\u7684\u7075\u6d3b\u6027\uff0c\u53ef\u7528\u4e8e\u8d1f\u8f7d\u5747\u8861\u7b49\u573a\u666f\u3002\u4f46\u5b9e\u9645\u90e8\u7f72\u4e2d\uff0cINT\u6570\u636e\u91c7\u96c6\u9762\u4e34\u6570\u636e\u566a\u58f0\uff08\u4e22\u5305\u3001\u5ef6\u8fdf\u4e71\u5e8f\u3001\u54c8\u5e0c\u51b2\u7a81\uff09\u4e0e\u7f51\u7edc\u52a8\u6001\u6027\uff08\u8def\u5f84\u8c03\u6574\u7b49\uff09\u7684\u6311\u6218\u3002\u5f53\u524d\u5229\u7528\u795e\u7ecf\u7f51\u7edc\u5206\u6790INT\u6570\u636e\u7684\u7814\u7a76\uff0c\u672a\u5145\u5206\u8003\u8651\u4ea4\u6362\u673a\u8d44\u6e90\u9650\u5236\u53ca\u4e0a\u8ff0\u95ee\u9898\uff0c\u672c\u6587\u63d0\u51faPlanner\u751f\u6210\u5f0f\u56fe\u5b66\u4e60\u6846\u67b6\uff0c\u901a\u8fc7\u6a21\u5757\u5316\u591a\u7c92\u5ea6\u7279\u5f81\u91c7\u96c6\u3001\u52a8\u6001\u56fe\u6784\u5efa\u7b49\u673a\u5236\u6765\u89e3\u51b3\u95ee\u9898\u3002<\/p>\n\n\n\n

      \u6838\u5fc3\u601d\u60f3\u4e0e\u8d21\u732e<\/strong><\/p>\n\n\n\n

      Planner \u5c1d\u8bd5\u5728\u9002\u914d\u53ef\u7f16\u7a0b\u4ea4\u6362\u673a\u7684\u7279\u6027\u4e0b\uff0c\u6784\u5efa\u6a21\u5757\u5316\u7684INT\u6846\u67b6\u4e0e\u6df7\u5408\u5b66\u4e60\u6a21\u578b\u3002\u5177\u4f53\u800c\u8a00\uff0c\u652f\u6301\u591a\u7c92\u5ea6\u7279\u5f81\u6765\u91c7\u96c6\u8bbe\u5907\u7ea7\u3001\u6d41\u7ea7\u3001\u5305\u7ea7\u4e09\u7c7b\u7279\u5f81\uff0c\u540c\u65f6\uff0c\u57fa\u4e8eINT\u8def\u5f84\u6784\u5efa\u52a8\u6001\u56fe\uff0c\u901a\u8fc7\u8282\u70b9\u7279\u5f81\u805a\u5408\u6765\u9002\u914d\u8def\u5f84\u7684\u53d8\u5316\uff0c\u4f5c\u8005\u5728\u57fa\u4e8eTofino\u53ef\u7f16\u7a0b\u4ea4\u6362\u673a\u548cBMv2\u6784\u5efa\u4e86testbed\u9a8c\u8bc1\uff0c\u5728\u653b\u51fb\u68c0\u6d4b\u3001\u5ef6\u8fdf\u9884\u6d4b\u4efb\u52a1\u4e0a\u4f18\u4e8e\u591a\u79cdbaseline\u3002<\/p>\n\n\n\n

      \u5b9e\u9a8c\u8bc4\u4f30<\/strong><\/p>\n\n\n\n

      \u5b9e\u9a8c\u5728\u653b\u51fb\u68c0\u6d4b\u3001\u5ef6\u8fdf\u9884\u6d4b\u4e24\u4e2a\u573a\u666f\u4e0b\u8bc4\u4f30\u6a21\u578b\u548cbaseline\u3002\u6a21\u62df\u4e86\u9ad8\u65af\u566a\u58f0\u3001\u63a9\u7801\u566a\u58f0\u4ee5\u53ca\u7279\u5f81\u653b\u51fb\u7b49\u591a\u79cd\u5e72\u6270\u3002\u5728\u6d4b\u8bd5\u573a\u666f\u4e2d\uff0c\u51c6\u786e\u7387\u548c\u53ec\u56de\u7387\u663e\u8457\u4f18\u4e8eVAE\u7b49\u591a\u4e2abaseline\uff0c\u5728\u52a8\u6001\u8def\u5f84\u53d8\u5316\u573a\u666f\u4e0b\u8868\u73b0\u6700\u4f18\u3002\u6d88\u878d\u5b9e\u9a8c\u8868\u660e\uff0c\u5404\u7ec4\u4ef6\u5747\u5177\u6709\u5fc5\u8981\u6027\u3002\u4e14\u786c\u4ef6\u8d44\u6e90\u548c\u65f6\u95f4\u8d44\u6e90\u7684\u6d88\u8017\u5747\u5728\u53ef\u63a5\u53d7\u8303\u56f4\u5185\u3002<\/p>\n\n\n\n

      \u95ee\u7b54\uff1a<\/strong><\/p>\n\n\n\n

        \n
      1. \u6a21\u578b\u662f\u600e\u4e48\u8bad\u7ec3\u7684\uff1f\u6a21\u578b\u5982\u4f55\u9002\u914d\u591a\u79cd\u62d3\u6251\uff1f<\/li>\n<\/ol>\n\n\n\n

        A: \u8bad\u7ec3\u6570\u636e\u6765\u81ea\u5b9e\u9a8c\u4e2d\u6536\u96c6\u7684\u6570\u636e\u96c6\uff0c\u6211\u4eec\u5c06\u6570\u636e\u96c6\u6309\u7167\u8bad\u7ec3\u96c6\u3001\u6d4b\u8bd5\u96c6\u8fdb\u884c\u5212\u5206\u3002\u5728\u6a21\u578b\u6570\u636e\u6536\u96c6\u9636\u6bb5\u8003\u8651\u4e86\u591a\u6837\u6027\u3002<\/p>\n\n\n\n

        \u4e2a\u4eba\u8bc4\u4ef7\uff1a<\/strong><\/p>\n\n\n\n

        Planner\u5f25\u5408\u4e86INT\u6570\u636e\u8d28\u91cf\u4e0e\u6a21\u578b\u9700\u6c42\u7684\u4e00\u4e2a\u5dee\u8ddd\uff0c\u5728\u653b\u51fb\u68c0\u6d4b\u3001\u5ef6\u8fdf\u9884\u6d4b\u573a\u666f\u4e0b\u6027\u80fd\u8868\u73b0\u548c\u9c81\u68d2\u6027\u5747\u4f73\u3002\u9700\u8981\u63a2\u7a76\u7684\u662f\uff0c\u5982\u4f55\u6269\u5c55\u81f3\u5176\u4ed6\u4efb\u52a1\uff0c\u8fdb\u4e00\u6b65\u62d3\u5bbd\u5e94\u7528\u8303\u56f4\u3002<\/p>\n\n\n\n

        Modeling Flow-level Traffic Demand for Network Performance Evaluation and Optimization<\/p>\n\n\n\n

        Sijiang Huang\u2021, Xiaohui Xie\u2021\u2217, Mowei Wang\u2020, Lingfeng Peng\u2021, Cong Li\u00a7, Yong Zhang\u00a7, Yingjie Qin\u00a7, Liang Zhang\u2020, Yong Cui\u2021\u2217 <\/p>\n\n\n\n

         \u2021 Tsinghua University, Beijing, China <\/p>\n\n\n\n

         \u2020Huawei Technologies, Beijing, China <\/p>\n\n\n\n

         \u00a7Industrial and Commercial Bank of China, Beijing, China<\/p>\n\n\n\n

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

        \u4ecb\u7ecd<\/strong><\/p>\n\n\n\n

        \u6d41\u91cf\u5efa\u6a21\u4e0e\u751f\u6210\u662fTE\u7684\u6838\u5fc3\u9700\u6c42\uff0c\u6d41\u7ea7\u522b\u6a21\u578b\uff08\u4ee5\u4e94\u5143\u7ec4\u8868\u5f81\uff09\u56e0\u7cbe\u51c6\u6355\u6349\u5e94\u7528\u9700\u6c42\uff0c\u5728\u7f51\u7edc\u8bc4\u4f30\u3001\u8bbe\u8ba1\u9a8c\u8bc1\u7b49\u573a\u666f\u4e2d\u6709\u91cd\u8981\u4ef7\u503c\u3002\u4f46\u4e3b\u6d41\u65b9\u6cd5\u4f9d\u8d56\u6d41\u91cf\u5206\u5e03\u540c\u8d28\u5316\u7b49\u4e0e\u5b9e\u9645\u4e0d\u7b26\u7684\u7b80\u5316\u5047\u8bbe\uff0c\u5bfc\u81f4\u751f\u6210\u7684\u6d41\u91cf\u4fdd\u771f\u5ea6\u4f4e\uff0c\u4e25\u91cd\u5f71\u54cd\u8bc4\u4f30\u4f18\u5316\u7684\u51c6\u786e\u6027\u3002\u672c\u65b9\u6848\u901a\u8fc7\u5206\u6790\u67d0\u5927\u578b\u4f01\u4e1a\u7684\u751f\u4ea7\u6d41\u91cf\uff0c\u63d0\u51fa Encore \u6846\u67b6\uff0c\u4ee5\u5206\u800c\u6cbb\u4e4b\u7684\u7b56\u7565\uff0c\u901a\u8fc7CVAE\u6355\u6349\u5206\u5e03\u5f02\u8d28\u6027\u3001\u6761\u4ef6GRU\u6355\u6349\u65f6\u95f4\u52a8\u6001\u6027\uff0c\u4ece\u800c\u751f\u6210\u9ad8\u4fdd\u771f\u5408\u6210\u6d41\u91cf\u8f68\u8ff9\u3002<\/p>\n\n\n\n

        \u6838\u5fc3\u601d\u60f3\u4e0e\u8d21\u732e<\/strong><\/p>\n\n\n\n

        Encore\u9488\u5bf9\u6d41\u7ea7\u522b\u6d41\u91cf\u7684\u5206\u5e03\u5f02\u8d28\u6027\u4e0e\u5e8f\u5217\u65f6\u95f4\u4f9d\u8d56\u6027\u4e24\u4e2a\u7279\u5f81\u5206\u522b\u5efa\u6a21\u3002\u5728\u5206\u5e03\u4e0a\uff0c\u5229\u7528CVAE\u505a\u7279\u5f81\u63d0\u53d6\uff0c\u540c\u65f6\u5f15\u5165\u52a8\u6001\u5e73\u8861\u4ee5\u53caKL\u6563\u5ea6\u7684loss\uff0c\u89e3\u51b3\u65f6\u95f4\u5206\u5e03\u5dee\u5f02\u95ee\u9898\uff1b\u5728\u5e8f\u5217\u4e0a\uff0c\u57fa\u4e8eGRU\u6a21\u578b\u5904\u7406\u65f6\u5e8f\u6570\u636e\uff0c\u5c06\u6d41\u5927\u5c0f\u4e0e\u5230\u8fbe\u65f6\u95f4\u7ec4\u6210\u7684 \u201c\u6d41\u5bf9\u201d \u4f5c\u4e3a\u57fa\u672c\u5efa\u6a21\u5355\u5143\u8fdb\u884c\u8bad\u7ec3\u3002\u6b64\u5916\uff0cEncore \u5728\u8bbe\u8ba1\u4e2d\u4fdd\u7559\u4e86\u9690\u79c1\u4fdd\u62a4\uff0c\u901a\u8fc7\u6a21\u578b\u53c2\u6570\u5b9e\u73b0\u6d41\u91cf\u751f\u6210\uff0c\u907f\u514d\u4fe1\u606f\u6cc4\u9732\u3002<\/p>\n\n\n\n

        \u5b9e\u9a8c\u8bc4\u4f30<\/strong><\/p>\n\n\n\n

        \u6570\u636e\u96c6\u6765\u81ea\u67d0\u5927\u578b\u5546\u4e1a\u94f6\u884c\u7684\u9aa8\u5e72\u7f51\u51fa\u53e3\u6d41\u91cf\u6570\u636e\uff0c\u4ee5\u53ca\u7528\u4e8e\u9a8c\u8bc1\u6a21\u578b\u6548\u679c\u7684\u516c\u5f00\u6570\u636e\u96c6\u3002\u5bf9\u6bd4\u4e86\u5206\u5c42\u8d1d\u53f6\u65af\u7b49\u4e3b\u6d41\u6a21\u578b\u3002\u672c\u65b9\u6848\u5728\u5206\u5e03\u5efa\u6a21\u4e0a\uff0c\u51c6\u786e\u7387\u548c\u8986\u76d6\u7387\u5747\u5927\u5e45\u4f18\u4e8e\u73b0\u6709\u65b9\u6cd5\u3002\u5728\u5e8f\u5217\u5efa\u6a21\u4e2d\uff0c\u9664\u5c11\u6570\u60c5\u51b5\u6301\u5e73\u5916\uff0c\u5927\u591a\u6570\u60c5\u51b5\u6307\u6807\u5747\u4f18\u4e8e\u73b0\u6709baseline\u3002\u540c\u65f6\uff0c\u672c\u65b9\u6848\u7684\u8bad\u7ec3\u6548\u7387\u9ad8\uff0c\u6210\u672c\u9002\u4e2d\u3002\u751f\u6210\u6548\u7387\u8fdc\u4f18\u4e8e\u73b0\u6709\u7684\u7f51\u7edc\u6a21\u62df\u5668\u3002<\/p>\n\n\n\n

        \u95ee\u7b54\uff1a<\/strong><\/p>\n\n\n\n

          \n
        1. \u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u4f7f\u7528\u7684\u6570\u636e\u96c6\u662f\u4ec0\u4e48\uff1f\u662f\u79c1\u6709\u6570\u636e\u5417\uff1f<\/li>\n<\/ol>\n\n\n\n

          A: \u6570\u636e\u96c6\u6765\u81ea\u4e24\u4e2a\u6765\u6e90\uff0c\u4e00\u662f\u57fa\u4e8e\u67d0\u4f01\u4e1a\u7684\u79c1\u6709\u6570\u636e\u96c6\uff0c\u4e8c\u662f\u4e92\u8054\u7f51\u7684\u516c\u5f00\u6570\u636e\u96c6\u3002<\/p>\n\n\n\n

          \u4e2a\u4eba\u8bc4\u4ef7\uff1a<\/strong><\/p>\n\n\n\n

          Encore\u5229\u7528\u6a21\u578b\u80fd\u529b\uff0c\u63d0\u4f9b\u4e86 \u201c\u9ad8\u4fdd\u771f\u4e14\u6613\u7528\u201d \u7684\u6d41\u91cf\u751f\u6210\u5de5\u5177\u3002\u5b9e\u9645\u4e2d\u53ef\u4ee5\u8fdb\u4e00\u6b65\u63a2\u8ba8\uff0c\u9664\u4e86\u652f\u6491\u7f51\u7edc\u6027\u80fd\u8bc4\u4f30\u4e0e\u4f18\u5316\u8fd9\u4e00\u4efb\u52a1\u5916\uff0c\u5982\u4f55\u6269\u5c55\u81f3\u5176\u4ed6\u573a\u666f\uff0c\u4f8b\u5982\u751f\u6210\u653b\u51fb\u6d41\u91cf\u3001\u65b0\u534f\u8bae\u9a8c\u8bc1\u7b49\u4f9b\u591a\u6837\u6d41\u91cf\u8f93\u5165\u3002\u501f\u52a9 \u201c\u6d41\u91cf\u5927\u6a21\u578b\u201d \u6765\u6784\u5efa\u901a\u7528\u5efa\u6a21\u6846\u67b6\u53ef\u80fd\u6210\u4e3a\u672a\u6765\u8fd9\u4e00\u9886\u57df\u7684\u8d8b\u52bf\u3002<\/p>\n\n\n\n

          DANCE<\/p>\n\n\n\n

              \u5317\u4eac\u90ae\u7535\u5927\u5b66 DANCE Group\uff0c\u5373\u7f51\u7edc\u4e0e\u6570\u636e\u878d\u5408 (Data And Network CoordinatE) \u8bfe\u9898\u7ec4\uff0c\u6bcf\u6708\u8bba\u6587\u5206\u4eab\u3002<\/p>\n\n\n\n

          \u8bfe\u9898\u7ec4\u8be6\u7ec6\u4fe1\u606f\uff1a<\/p>\n\n\n\n

          http:\/\/yuchaozhang.weebly.com\/<\/p>\n\n\n\n

          \u62db\u751f\u4fe1\u606f (\u8ba1\u7b97\u673a17\u7ec4\u3001\u8f6f\u4ef606\u7ec4)\uff1a<\/p>\n\n\n\n

          \nhttps:\/\/scs.bupt.edu.cn\/info\/1020\/2675.htm\n<\/div><\/figure>\n\n\n\n

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

          \u7b2c 33 \u5c4a IEEE ICNP \u4e8e 2025 \u5e74 9 \u6708\u5728\u97e9\u56fd\u9996\u5c14\u53ec\u5f00\uff0c\u672c\u6b21\u5927\u4f1a\u5171\u8ba1 238 \u7bc7\u6295\u7a3f\uff0c\u5f55\u7528 60 \u7bc7\uff08\u5f55\u7528\u7387\u4e3a 25%\uff09\u3002 \u300aMamba4Net: \u84b8\u998f\u5f0f\u6df7\u5408Mamba\u5927\u8bed\u8a00\u6a21\u578b\u5728\u7f51\u7edc\u9886\u57df\u7684\u5e94\u7528\u300b\u8bba\u6587\u4ecb\u7ecd \u300aMamba4Net: Distilled Hybrid Mamba Large Language Models For Networking\u300b\u662f\u4e00\u7bc7\u53d1\u8868\u4e8e2025\u5e74IEEE\u7b2c33\u5c4a\u56fd\u9645\u7f51\u7edc\u534f\u8bae\u4f1a\u8bae\uff08ICNP\uff09\u7684\u8bba\u6587\uff0c\u5317\u4eac\u5927\u5b66\u6df1\u5733\u7814\u7a76\u751f\u9662ICNLab\u3002\u8be5\u8bba … \u7ee7\u7eed\u9605\u8bfb →<\/span><\/a><\/p>\n","protected":false},"author":2,"featured_media":524,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[17,23],"tags":[],"class_list":["post-507","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-17","category-23"],"_links":{"self":[{"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/posts\/507","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=507"}],"version-history":[{"count":3,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/posts\/507\/revisions"}],"predecessor-version":[{"id":525,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/posts\/507\/revisions\/525"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=\/wp\/v2\/media\/524"}],"wp:attachment":[{"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=507"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=507"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.ndnlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=507"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}