{"id":25481,"date":"2021-04-11T19:48:11","date_gmt":"2021-04-11T11:48:11","guid":{"rendered":"https:\/\/swarma.org\/?p=25481"},"modified":"2021-04-11T19:48:11","modified_gmt":"2021-04-11T11:48:11","slug":"%e4%bb%80%e4%b9%88%e6%98%af%e4%bb%80%e4%b9%88%e6%98%af%e8%b4%9d%e5%8f%b6%e6%96%af%e7%bd%91%e7%bb%9c%ef%bc%9a%e5%9b%be%e6%a8%a1%e5%9e%8b%e3%80%81%e6%8e%a8%e7%90%86%e4%b8%8e%e5%ad%a6%e4%b9%a0","status":"publish","type":"post","link":"https:\/\/swarma.org\/?p=25481","title":{"rendered":"\u4ec0\u4e48\u662f\u4ec0\u4e48\u662f\u8d1d\u53f6\u65af\u7f51\u7edc\uff1a\u56fe\u6a21\u578b\u3001\u63a8\u7406\u4e0e\u5b66\u4e60 | \u96c6\u667a\u767e\u79d1"},"content":{"rendered":"
\n

<\/p>\n


<\/section>\n
\u201c\u96c6\u667a\u767e\u79d1\u7cbe\u9009\u201d\u662f\u4e00\u4e2a\u957f\u671f\u4e13\u680f\uff0c\u6301\u7eed\u4e3a\u5927\u5bb6\u63a8\u9001\u590d\u6742\u6027\u79d1\u5b66\u76f8\u5173\u7684\u57fa\u672c\u6982\u5ff5\u548c\u8d44\u6e90\u4fe1\u606f\u3002\u4f5c\u4e3a\u96c6\u667a\u4ff1\u4e50\u90e8\u7684\u5f00\u6e90\u79d1\u5b66\u9879\u76ee\uff0c\u96c6\u667a\u767e\u79d1\u5e0c\u671b\u6253\u9020\u590d\u6742\u6027\u79d1\u5b66\u9886\u57df\u6700\u5168\u9762\u7684\u767e\u79d1\u5168\u4e66\uff0c\u6b22\u8fce\u5bf9\u590d\u6742\u6027\u79d1\u5b66\u611f\u5174\u8da3\u3001\u70ed\u7231\u77e5\u8bc6\u6574\u7406\u548c\u5206\u4eab\u7684\u670b\u53cb\u52a0\u5165\uff01
<\/span><\/section>\n

<\/section>\n
\u672c\u6587\u662f\u5bf9\u96c6\u667a\u767e\u79d1\u4e2d\u201c\u8d1d\u53f6\u65af\u7f51\u7edc\u201d\u8bcd\u6761\u90e8\u5206\u5185\u5bb9\u7684\u6458\u5f55\uff0c\u53c2\u8003\u8d44\u6599\u53ca\u76f8\u5173\u8bcd\u6761\u8bf7\u53c2\u9605\u767e\u79d1\u8bcd\u6761\u539f\u6587\u3002<\/span><\/section>\n

<\/section>\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n

\u672c\u8bcd\u6761\u7531\u96c6\u667a\u4ff1\u4e50\u90e8\u4f17\u5305\u751f\u4ea7\uff0c\u96be\u514d\u5b58\u5728\u7eb0\u6f0f\u548c\u95ee\u9898\uff0c\u6b22\u8fce\u5927\u5bb6\u7559\u8a00\u53cd\u9988\u6216\u8005\u524d\u5f80\u5bf9\u5e94\u7684\u767e\u79d1\u8bcd\u6761\u9875\u9762\u8fdb\u884c\u4fee\u6539\uff0c\u4e00\u7ecf\u4fee\u6539\uff0c\u53ef\u4ee5\u83b7\u5f97\u5bf9\u5e94\u7684\u79ef\u5206\u5956\u52b1\u5662\uff01<\/em><\/span><\/p>\n


<\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n

\u76ee\u5f55<\/strong><\/p>\n


<\/p>\n

\n
\u4e00\u3001\u56fe\u6a21\u578b<\/span><\/strong><\/section>\n
\u4e8c\u3001\u4e3e\u4f8b<\/span><\/strong><\/section>\n
\u4e09\u3001\u96c6\u667a\u767e\u79d1\u8bcd\u6761\u5fd7\u613f\u8005\u62db\u52df<\/strong><\/strong><\/strong><\/span><\/strong><\/span><\/strong><\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n

<\/section>\n
<\/section>\n

\u8d1d\u53f6\u65af\u7f51\u7edc Bayesian network\uff08BN\uff09\u3001\u4fe1\u5ff5\u7f51\u7edc belief network\u3001\u51b3\u7b56\u7f51\u7edc decision network\u3001\u8d1d\u53f6\u65af\u6a21\u578b Bayes(ian) model\u6216\u6982\u7387\u6709\u5411\u65e0\u73af\u56fe\u6a21\u578b probabilistic directed acyclic graphical model\u662f\u4e00\u79cd\u6982\u7387\u56fe\u6a21\u578b\uff08\u4e00\u79cd\u7edf\u8ba1\u6a21\u578b\uff09\uff0c\u5b83\u901a\u8fc7\u6709\u5411\u65e0\u73af\u56fe\u65e0\u73af\u56fe directed acyclic graph\uff08DAG\uff09\u8868\u793a\u4e00\u7ec4\u968f\u673a\u53d8\u91cf\u53ca\u5176\u6761\u4ef6\u4f9d\u8d56\u5173\u7cfb\u3002\u8d1d\u53f6\u65af\u7f51\u7edc\u662f\u4e00\u79cd\u7406\u60f3\u7684\u5206\u6790\u5de5\u5177\uff0c\u7528\u6765\u9884\u6d4b\u4e00\u4e2a\u4e8b\u4ef6\u7684\u53d1\u751f\u662f\u7531\u5df2\u77e5\u539f\u56e0\u4e2d\u7684\u54ea\u4e00\u4e2a\uff08\u4e9b\uff09\u5f15\u8d77\u7684\u3002\u4f8b\u5982\uff0c\u8d1d\u53f6\u65af\u7f51\u7edc\u53ef\u4ee5\u8868\u793a\u75be\u75c5\u548c\u75c7\u72b6\u4e4b\u95f4\u7684\u6982\u7387\u5173\u7cfb\u3002\u5728\u7ed9\u5b9a\u75c7\u72b6\u7684\u60c5\u51b5\u4e0b\uff0c\u8be5\u7f51\u7edc\u53ef\u7528\u4e8e\u8ba1\u7b97\u5404\u79cd\u75be\u75c5\u51fa\u73b0\u7684\u6982\u7387\u3002<\/span>
<\/p>\n


\u8d1d\u53f6\u65af\u7f51\u7edc\u6709\u591a\u79cd\u53d8\u4f53\u3002\u7528\u6765\u5efa\u6a21\u5e8f\u5217\u53d8\u91cf\uff08\u4f8b\u5982\uff0c\u8bed\u97f3\u4fe1\u53f7\u6216\u86cb\u767d\u8d28\u5e8f\u5217\uff09\u7684\u8d1d\u53f6\u65af\u7f51\u7edc\u88ab\u79f0\u4e3a\u52a8\u6001\u8d1d\u53f6\u65af\u7f51\u7edc dynamic Bayesian network\u3002\u8d1d\u53f6\u65af\u7f51\u7edc\u53ef\u4ee5\u8fdb\u4e00\u6b65\u6269\u5c55\uff0c\u7528\u6765\u8868\u793a\u548c\u89e3\u51b3\u5728\u4e0d\u786e\u5b9a\u6027\u56e0\u7d20\u4e0b\u7684\u51b3\u7b56\u95ee\u9898\uff0c\u8fd9\u79cd\u6269\u5c55\u79f0\u4e3a\u5f71\u54cd\u56fe influence diagram\u3002\u73b0\u5728\u5df2\u6709\u9ad8\u6548\u7684\u7b97\u6cd5\u5b66\u4e60\u51fa\u8d1d\u53f6\u65af\u7f51\u7edc\u7684\u7ed3\u6784\uff0c\u5e76\u901a\u8fc7\u8d1d\u53f6\u65af\u7f51\u7edc\u505a\u63a8\u7406\u3002<\/span><\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n


<\/p>\n


<\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n

<\/section>\n

<\/section>\n
\n
<\/svg><\/section>\n<\/section>\n<\/section>\n
\n

\u56fe\u6a21\u578b<\/span><\/strong><\/strong><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n
<\/strong>
<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n


<\/p>\n

\u5728\u5f62\u5f0f\u4e0a\uff0c\u8d1d\u53f6\u65af\u7f51\u7edc\u662f\u6709\u5411\u65e0\u73af\u56fe Directed acyclic graph\uff08DAG\uff09\uff0c\u5176\u8282\u70b9\u8868\u793a\u968f\u673a\u53d8\u91cf\uff08\u5176\u6982\u7387\u4e3a\u8d1d\u53f6\u65af\u6982\u7387\uff09\uff1b\u5b83\u4eec\u53ef\u4ee5\u662f\u53ef\u89c2\u6d4b\u91cf\u53d8\u91cf\u3001\u9690\u53d8\u91cf\u3001\u672a\u77e5\u53c2\u6570\u6216\u5047\u8bbe\u3002\u56fe\u4e2d\u7684\u8fb9\u8868\u793a\u6761\u4ef6\u4f9d\u8d56\uff1b\u672a\u8fde\u63a5\uff08\u6ca1\u6709\u8def\u5f84\u8fde\u63a5\u4e00\u4e2a\u8282\u70b9\u5230\u53e6\u4e00\u4e2a\u8282\u70b9\uff09\u7684\u8282\u70b9\u8868\u793a\u5f7c\u6b64\u4e4b\u95f4\u6761\u4ef6\u72ec\u7acb\u3002\u6bcf\u4e2a\u8282\u70b9\u90fd\u4e0e\u4e00\u4e2a\u6982\u7387\u51fd\u6570 Probability function \u8282\u70b9\u76f8\u5173\u8054\uff0c\u8be5\u51fd\u6570\u628a\u6240\u6709\u7236\u8282\u70b9\u4ee3\u8868\u7684\u53d8\u91cf\u503c\u4f5c\u4e3a\u8f93\u5165\uff0c\u5e76\u7ed9\u51fa\u8be5\u8282\u70b9\u8868\u793a\u7684\u968f\u673a\u53d8\u91cf\u7684\u6982\u7387\uff08\u6216\u6982\u7387\u5206\u5e03\uff09\u3002\u4f8b\u5982\uff0c\u5982\u679c m \u7236\u8282\u70b9\u8868\u793a m \u5e03\u5c14\u53d8\u91cf\uff0c\u90a3\u4e48\u6982\u7387\u51fd\u6570\u53ef\u4ee5\u7528\u4e00\u4e2a\u5305\u542b 2^m\u884c\u7684\u8868\u683c\u8868\u793a\uff0c\u6bcf\u4e00\u884c\u4ee3\u8868\u4e00\u79cd\uff08\u7236\u8282\u70b9\uff09\u53d8\u91cf\u503c\u7684\u7ec4\u5408\uff0c\u4ee5\u53ca\u5bf9\u5e94\u7684\u5b50\u8282\u70b9\u53d8\u91cf\u7684\u6982\u7387\u503c\u3002\u7c7b\u4f3c\u7684\u60f3\u6cd5\u53ef\u4ee5\u5e94\u7528\u4e8e\u6709\u73af\u65e0\u5411\u56fe\uff0c\u5982<\/span>\u9a6c\u5c14\u53ef\u592b\u7f51\u7edc Markov network<\/span>\u3002<\/span><\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n


<\/p>\n


<\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n

<\/section>\n

<\/section>\n
\n
<\/svg><\/section>\n<\/section>\n<\/section>\n
\n

\u4e3e\u4f8b<\/span><\/strong><\/strong><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n
<\/strong>
<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n


<\/p>\n

\u8349\u5730\u53d8\u5f97\u6e7f\u6da6\uff0c\u53ef\u80fd\u6709\u4e24\u79cd\u539f\u56e0\uff1a\u4e3b\u52a8\u6d12\u6c34\u6216\u8005\u4e0b\u96e8\u3002\u96e8\u5bf9\u6d12\u6c34\u8f66\u7684\u4f7f\u7528\u6709\u76f4\u63a5\u7684\u5f71\u54cd\uff08\u4e5f\u5c31\u662f\u8bf4\uff0c\u5f53\u4e0b\u96e8\u65f6\uff0c\u6d12\u6c34\u8f66\u901a\u5e38\u662f\u4e0d\u5de5\u4f5c\u7684\uff09\u3002\u8fd9\u79cd\u60c5\u51b5\u53ef\u4ee5\u7528\u8d1d\u53f6\u65af\u7f51\u7edc\u6765\u6a21\u62df\uff08\u5982\u53f3\u56fe\u6240\u793a\uff09\u3002\u6bcf\u4e2a\u53d8\u91cf\u6709\u4e24\u4e2a\u53ef\u80fd\u7684\u503c\uff0cT (\u8868\u793a\u771f)\u548c F (\u8868\u793a\u5047)\u3002<\/span><\/p>\n


\u5bf9\u5e94\u7684\u8054\u5408\u6982\u7387\u51fd\u6570 Joint probability distribution\u662f:<\/span><\/p>\n

<\/p>\n

\u5176\u4e2dG\u8868\u793a\u201c\u8349\u5730\u6e7f\u4e86\uff08\u662f\/\u5426\uff09\u201d\uff0cS\u8868\u793a\u201c\u6d12\u6c34\u5668\u6253\u5f00\uff08\u662f\/\u5426\uff09\u201d\uff0cR\u8868\u793a\u201c\u4e0b\u96e8\uff08\u662f\/\u5426\uff09\u201d\u3002
<\/span><\/p>\n


\u8fd9\u4e2a\u6a21\u578b\u53ef\u4ee5\u56de\u7b54\u5728\u7ed9\u5b9a\u4e00\u4e2a\u7ed3\u679c\u7684\u60c5\u51b5\u4e0b\u4e00\u4e2a\u539f\u56e0\u662f\u5426\u5b58\u5728\u7684\u95ee\u9898\uff0c\u6bd4\u5982\u201c\u7ed9\u5b9a\u8349\u662f\u6e7f\u7684\uff0c\u4e0b\u96e8\u7684\u6982\u7387\u662f\u591a\u5c11? \u201d\u901a\u8fc7\u4f7f\u7528\u6761\u4ef6\u6982\u7387\u516c\u5f0f\u5e76\u5bf9\u6240\u6709<\/span>\u5e72\u6270\u53d8\u91cf nuisance variable<\/span>\u7684\u6c42\u548c:<\/span><\/p>\n

<\/p>\n


<\/p>\n

\u5c55\u5f00\u6982\u7387\u51fd\u6570 Pr (G,S,R) \uff0c\u5e76\u4f7f\u7528\u56fe\u4e2d\u5217\u51fa\u7684\u6761\u4ef6\u6982\u7387\uff0c\u6211\u4eec\u53ef\u4ee5\u7b97\u51fa\u5206\u5b50\u548c\u5206\u6bcd\u4e2d\u7684\u5404\u4e2a\u9879\u3002\u6bd4\u5982\u8bf4,
<\/span><\/p>\n

<\/p>\n

\u7b97\u51fa\u6765\u7684\u7ed3\u679c\u662f<\/span>
<\/p>\n

<\/section>\n

\u8fd9\u4e2a\u6a21\u578b\u8fd8\u56de\u7b54\u5e72\u9884\u6027\u7684\u95ee\u9898\uff0c\u6bd4\u5982\u201c\u73b0\u5728\u6211\u4eec\u628a\u8349\u5f04\u6e7f\u4e86\uff0c\u90a3\u4e48\u4e0b\u96e8\u7684\u53ef\u80fd\u6027\u6709\u591a\u5927? \u201d\u7b54\u6848\u53d6\u51b3\u4e8e\u5e72\u9884\u540e\u7684\u8054\u5408\u5206\u5e03\u51fd\u6570\uff1a<\/span>
<\/p>\n

<\/p>\n

\u8be5\u5206\u5e03\u901a\u8fc7\u4ece\u5e72\u9884\u524d\u7684\u5206\u5e03\u4e2d\u53bb\u9664\u56e0\u5b50Pr (G|S,R) \u5f97\u5230\uff0c\u5176\u4e2ddo\u7b97\u5b50\u5f3a\u884c\u4f7f G \u7684\u503c\u4e3a\u771f\u3002\u6f14\u7b97\u540e\u53ef\u77e5\u4e0b\u96e8\u7684\u53ef\u80fd\u6027\u4e0d\u53d7\u6b64\u5e72\u9884\u7684\u5f71\u54cd:<\/span><\/p>\n

<\/p>\n

\u73b0\u5728\u518d\u9884\u6d4b\u5f00\u542f\u6d12\u6c34\u88c5\u7f6e\u7684\u5f71\u54cd:<\/span><\/p>\n

<\/p>\n

\u79fb\u9664Pr (S=T|R) \u8fd9\u4e2a\u9879\u8868\u660e\u8fd9\u79cd\u884c\u4e3a\u5f71\u54cd\u7684\u662f\u8349\uff0c\u800c\u4e0d\u662f\u96e8\u3002<\/span><\/p>\n

\u8003\u8651\u5230\u672a\u89c2\u6d4b\u53d8\u91cf\uff0c\u8fd9\u4e9b\u9884\u6d4b\u53ef\u80fd\u5e76\u4e0d\u53ef\u884c\uff0c\u5c31\u50cf\u5927\u591a\u6570\u7b56\u7565\u8bc4\u4f30\u95ee\u9898\u4e00\u6837\u3002\u4f46\u662f\uff0c\u53ea\u8981\u6ee1\u8db3\u540e\u95e8\u51c6\u5219\uff0c\u4ecd\u7136\u53ef\u4ee5\u9884\u6d4bdo(x) \u7684\u6548\u679c\u3002<\/span>\u5982\u679c\u4e00\u7ec4\u89c2\u5bdf\u5230\u7684\u53d8\u91cfZ\u80fdd-\u5206\u9694(\u6216\u963b\u585e)<\/span>[3]<\/span>\u4ece X \u5230 Y \u7684\u6240\u6709”\u540e\u95e8\u8def\u5f84 back-door path\uff0c\u5219\u6709<\/span><\/p>\n

<\/p>\n

\u540e\u95e8\u8def\u5f84\u662f\u4e00\u6761\u7bad\u5934\u6307\u5411X\u7684\u8def\u5f84\u3002\u6ee1\u8db3\u540e\u95e8\u6807\u51c6\u7684\uff08\u89c2\u6d4b\u53d8\u91cf\uff09\u96c6\u5408\u79f0\u4e3a\u201c\u5145\u5206\u7684\u201d\u6216\u201c\u6709\u6548\u7684\u201d\u3002\u4f8b\u5982\uff0c\u96c6\u5408 Z = R \u80fd\u6709\u6548\u5730\u9884\u6d4b S = T \u5bf9 G \u7684\u5f71\u54cd\uff0c\u56e0\u4e3a R d-\u5206\u9694\u4e86(\u4ec5\u6709\u7684)\u540e\u95e8\u8def\u5f84 S \u2190 R \u2192 G\u3002\u4f46\u662f\uff0c\u5982\u679c S \u6ca1\u6709\u88ab\u89c2\u6d4b\u5230\uff0c\u6ca1\u6709\u5176\u4ed6\u89c2\u6d4b\u53d8\u91cf\u96c6\u5408\u6765 d-\u5206\u9694\u8fd9\u6761\u8def\u5f84\uff0c\u90a3\u5c31\u4e0d\u80fd\u4ece\u89c2\u6d4b\u6570\u636e\u4e2d\u9884\u6d4b\u5230\u201c\u55b7\u5934\u88ab\u6253\u5f00\u201d(S = T)\u5bf9\u4e8e\u8349\u5730G\u7684\u5f71\u54cd\u3002\u5728\u8fd9\u79cd\u60c5\u51b5\u4e0b\uff0cP(G | do(S = T))\u5c31\u6ca1\u6709\u88ab\u201c\u8bc6\u522b\u201d\u3002\u8fd9\u53cd\u6620\u4e86\u4e00\u4e2a\u4e8b\u5b9e\uff1a\u5728\u7f3a\u4e4f\u5e72\u9884\u6027\u6570\u636e\u7684\u60c5\u51b5\u4e0b\u65e0\u6cd5\u786e\u8ba4\u89c2\u5bdf\u5230\u7684 S \u548c G \u4e4b\u95f4\u7684\u4f9d\u8d56\u5173\u7cfb\u662f\u4e0d\u662f\u4e00\u79cd\u56e0\u679c\u5173\u7cfb(\u53ef\u80fd\u7531\u5171\u540c\u539f\u56e0\u5f15\u8d77\u7684\u5f3a\u76f8\u5173\uff0c\u6bd4\u5982\u8f9b\u666e\u68ee\u6096\u8bba)\u3002<\/span><\/p>\n


\u4e3a\u4e86\u786e\u5b9a\u4e00\u4e2a\u56e0\u679c\u5173\u7cfb\u662f\u5426\u53ef\u4ee5\u4ece\u4e00\u4e2a\u542b\u6709\u672a\u89c2\u6d4b\u53d8\u91cf\u7684\u8d1d\u53f6\u65af\u7f51\u7edc\u4e2d\u8bc6\u522b\u51fa\u6765\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u201c do-\u6f14\u7b97\u201d<\/span>\u7684\u4e09\u4e2a\u89c4\u5219\u6765\u68c0\u9a8c\u662f\u5426\u6240\u6709\u7684 do \u9879\u90fd\u53ef\u4ee5\u4ece\u8fd9\u4e2a\u5173\u7cfb\u7684\u8868\u8fbe\u5f0f\u4e2d\u53bb\u6389\uff0c\u4ece\u800c\u786e\u8ba4\u6240\u9700\u7684\u91cf\u53ef\u4ee5\u4ece\u6570\u636e\u4e2d\u4f30\u8ba1\u51fa\u6765\u3002<\/span><\/p>\n


\u5982\u679c\u4f9d\u8d56\u5173\u7cfb\u5728\u8054\u5408\u5206\u5e03\u4e2d\u662f\u7a00\u758f\u7684\uff08\u53d8\u91cf\u95f4\u4f9d\u8d56\u8f83\u5c11\uff0c\u5373\u5bf9\u5e94\u7684\u56fe\u6a21\u578b\u91cc\u7684\u8fb9\u8f83\u5c11\uff09\uff0c\u90a3\u4e48\u76f8\u5bf9\u4e8e\u5b58\u50a8\u4e00\u5f20\u5b8c\u6574\u7684\u6982\u7387\u8868\uff0c\u4f7f\u7528\u8d1d\u53f6\u65af\u7f51\u7edc\u53ef\u4ee5\u8282\u7701\u76f8\u5f53\u591a\u7684\u5185\u5b58\u3002\u4f8b\u5982\uff0c\u5c0610\u4e2a\u4e8c\u503c\u53d8\u91cf\u7684\u6761\u4ef6\u6982\u7387\u5b58\u50a8\u4e3a\u4e00\u4e2a\u8868\u7684\uff0c\u9700\u8981\u5b58\u50a8 2^10=1024 \u4e2a\u503c\u3002\u800c\u5728\u6bcf\u4e2a\u53d8\u91cf\u6700\u591a\u4f9d\u8d563\u4e2a\u7236\u53d8\u91cf\u7684\u60c5\u51b5\u4e0b\uff0c\u4f7f\u7528\u8d1d\u53f6\u65af\u7f51\u7edc\u8868\u793a\u6700\u591a\u53ea\u5b58\u50a8 10\u00b72^3=80 \u4e2a\u503c\u3002<\/span><\/p>\n


<\/p>\n

\u76f8\u6bd4\u4e8e\u5b8c\u5168\u7248\u7684\u8054\u5408\u6982\u7387\u5206\u5e03\uff0c\u7406\u89e3\uff08\u4e00\u7ec4\u7a00\u758f\u7684\uff09\u76f4\u63a5\u7684\u53d8\u91cf\u95f4\u4f9d\u8d56\u5173\u7cfb\u548c\u5c40\u90e8\u7684\u6982\u7387\u5206\u5e03\u5bf9\u4e8e\u4eba\u7c7b\u6765\u8bf4\u8981\u66f4\u52a0\u76f4\u89c2\u6613\u61c2\u3002\u8fd9\u6b63\u662f\u8d1d\u53f6\u65af\u7f51\u7edc\u7684\u4e00\u4e2a\u4f18\u70b9\u3002<\/span><\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n


<\/p>\n


<\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n

<\/section>\n

<\/section>\n
\n
<\/svg><\/section>\n<\/section>\n<\/section>\n
\n

\u7f16\u8005\u63a8\u8350<\/span><\/strong><\/strong><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n
<\/strong>
<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n


<\/p>\n

\u56e0\u679c\u79d1\u5b66\u4e0e Causal AI \u8bfb\u4e66\u4f1a\u7b2c\u4e8c\u5b63<\/span><\/strong><\/span><\/section>\n
\u56e0\u679c\u63a8\u65ad\u4e0e\u673a\u5668\u5b66\u4e60\u9886\u57df\u7684\u7ed3\u5408\u5df2\u7ecf\u5438\u5f15\u4e86\u8d8a\u6765\u8d8a\u591a\u6765\u81ea\u5b66\u754c\u4e1a\u754c\u7684\u5173\u6ce8\uff0c\u4e3a\u6df1\u5165\u63a2\u8ba8\u3001\u666e\u53ca\u63a8\u5e7f\u56e0\u679c\u79d1\u5b66\u8bae\u9898\uff0c\u667a\u6e90\u793e\u533a\u643a\u624b\u96c6\u667a\u4ff1\u4e50\u90e8\u5c06\u4e3e\u529e\u7b2c\u4e8c\u5b63\u300c\u56e0\u679c\u79d1\u5b66\u4e0eCausalAI\u8bfb\u4e66\u4f1a\u300d\u3002\u672c\u5b63\u8bfb\u4e66\u4f1a\u7740\u529b\u4e8e\u5b9e\u64cd\u6027\u3001\u57fa\u7840\u6027\uff0c\u5c06\u5e26\u9886\u5927\u5bb6\u7cbe\u8bfb\u56e0\u679c\u79d1\u5b66\u65b9\u5411\u4e24\u672c\u975e\u5e38\u53d7\u5e7f\u6cdb\u8ba4\u53ef\u7684\u5165\u95e8\u6559\u6750\u3002<\/span><\/section>\n

<\/section>\n
1. Pearl, Judea, Madelyn Glymour, and Nicholas P. Jewell. Causal inference in statistics: A primer. John Wiley & Sons, 2016.\uff08\u672c\u4e66\u4e2d\u8bd1\u7248\u300a\u7edf\u8ba1\u56e0\u679c\u63a8\u7406\u5165\u95e8\uff08\u7ffb\u8bd1\u7248\uff09\u300b\u5df2\u7531\u9ad8\u7b49\u6559\u80b2\u51fa\u7248\u793e\u51fa\u7248\uff09<\/span><\/section>\n

<\/section>\n
2. Peters, Jonas, Dominik Janzing, and Bernhard Sch\u00f6lkopf. Elements of causal inference: foundations and learning algorithms. The MIT Press, 2017.<\/span><\/section>\n

<\/section>\n
\u8bfb\u4e66\u4f1a\u6bcf\u5468\u5c06\u8fdb\u884c\u76f4\u64ad\u8ba8\u8bba\uff0c\u8fdb\u884c\u95ee\u9898\u4ea4\u6d41\u3001\u91cd\u70b9\u6982\u5ff5\u5206\u4eab\u3001\u9605\u8bfb\u6982\u89c8\u548c\u7f16\u7a0b\u5b9e\u8df5\u5185\u5bb9\u5206\u6790\u3002\u975e\u5e38\u9002\u5408\u6709\u673a\u5668\u5b66\u4e60\u80cc\u666f\uff0c\u5e0c\u671b\u6df1\u5165\u5b66\u4e60\u56e0\u679c\u79d1\u5b66\u57fa\u7840\u77e5\u8bc6\u548c\u91cd\u8981\u6a21\u578b\u65b9\u6cd5\uff0c\u5bfb\u6c42\u89e3\u51b3\u76f8\u5173\u7814\u7a76\u95ee\u9898\u7684\u670b\u53cb\u53c2\u52a0\u3002\u8be6\u60c5\u53c2\u89c1\uff1a<\/span>\u8fde\u63a5\u7edf\u8ba1\u5b66\u3001\u673a\u5668\u5b66\u4e60\u4e0e\u81ea\u52a8\u63a8\u7406\u7684\u65b0\u5174\u4ea4\u53c9\u9886\u57df\u2014\u2014\u56e0\u679c\u79d1\u5b66\u8bfb\u4e66\u4f1a\u518d\u542f\u822a<\/a>\u3002<\/span><\/section>\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n


<\/p>\n


<\/p>\n

\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n

<\/section>\n

<\/section>\n
\n
<\/svg><\/section>\n<\/section>\n<\/section>\n
\n

\u767e\u79d1\u9879\u76ee\u5fd7\u613f\u8005\u62db\u52df<\/span><\/strong><\/strong><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n
<\/strong>
<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n


<\/p>\n

\u4f5c\u4e3a\u96c6\u667a\u767e\u79d1\u9879\u76ee\u56e2\u961f\u7684\u6210\u5458\uff0c\u672c\u6587\u5185\u5bb9<\/span>\u7531\u6c34\u6d41\u5fc3\u4e0d\u7ade<\/span><\/strong>\uff0c<\/span><\/strong>Ricky<\/span><\/strong>\uff0c<\/span><\/strong>\u8584\u8377<\/span><\/strong>\u53c2\u4e0e\u8d21\u732e\u3002\u6211\u4eec\u4e5f\u4e3a\u6bcf\u4f4d\u4f5c\u8005\u548c\u5fd7\u613f\u8005\u51c6\u5907\u4e86\u4e13\u5c5e\u7b80\u4ecb\u548c\u4e2a\u4eba\u96c6\u667a\u767e\u79d1\u4e3b\u9875\uff0c\u66f4\u591a\u4fe1\u606f\u53ef\u4ee5\u8bbf\u95ee\u5176\u96c6\u667a\u767e\u79d1\u4e2a\u4eba\u4e3b\u9875\u3002<\/span>
<\/p>\n


<\/p>\n

\u4ee5\u4e0a\u5185\u5bb9\u90fd\u662f\u6211\u4eec\u505a\u8fd9\u9879\u76ee\u7684\u8d77\u70b9\uff0c\u4f5c\u4e3a\u6765\u81ea\u4e0d\u540c\u5b66\u79d1\u548c\u9886\u57df\u7684\u5fd7\u613f\u8005\uff0c\u6211\u4eec\u5efa\u7acb\u8d77\u4e00\u4e2a\u6709\u6548\u7684\u767e\u79d1\u56e2\u961f\uff0c\u5206\u914d\u6709\u5ba1\u6821\u3001\u7ffb\u8bd1\u3001\u7f16\u8f91\u3001\u5ba3\u4f20\u7b49\u5de5\u4f5c\u3002\u6211\u4eec\u79c9\u6301\uff1a\u77e5\u8bc6\u4ece\u6211\u800c\u6765\uff0c\u95ee\u9898\u5230\u6211\u4e3a\u6b62\u7684\u4fe1\u5ff5\uff0c\u8ba4\u771f\u8d1f\u8d23\u7f16\u64b0\u6bcf\u4e00\u4e2a\u8bcd\u6761\u3002<\/span><\/section>\n


<\/p>\n


<\/p>\n

<\/p>\n


<\/p>\n

\u5728\u8fd9\u91cc\u4ece\u590d\u6742\u6027\u77e5\u8bc6\u51fa\u53d1\u4e0e\u4f19\u4f34\u540c\u884c\uff0c\u540c\u65f6\u6211\u4eec\u5e0c\u671b\u6709\u66f4\u591a\u5fd7\u613f\u8005\u52a0\u5165\u8fd9\u4e2a\u56e2\u961f\uff0c\u4f7f\u767e\u79d1\u8bcd\u6761\u5185\u5bb9\u5f97\u5230\u6269\u5145\uff0c\u5e76\u4e3a\u6bcf\u4f4d\u5fd7\u613f\u8005\u63d0\u4f9b\u76f8\u5e94\u5956\u52b1\u4e0e\u8d44\u6e90\uff0c\u5efa\u7acb\u4e2a\u4eba\u4e3b\u9875\u4e0e\u8d21\u732e\u8bb0\u5f55\uff0c\u4f7f\u5176\u80fd\u591f\u7ee7\u7eed\u63a2\u7d22\u590d\u6742\u4e16\u754c\u3002<\/span><\/p>\n


<\/p>\n

\u5982\u679c\u4f60\u6709\u610f\u53c2\u4e0e\u66f4\u52a0\u7cfb\u7edf\u7cbe\u7ec6\u7684\u5206\u5de5\uff0c\u626b\u63cf\u4e8c\u7ef4\u7801\u586b\u5199\u62a5\u540d\u8868\uff0c\u6211\u4eec\u671f\u5f85\u4f60\u7684\u52a0\u5165\uff01<\/span><\/p>\n


<\/p>\n


<\/p>\n

<\/section>\n
\u96c6\u667a\u767e\u79d1\u62a5\u540d\u8868<\/span><\/section>\n


<\/p>\n

\n
\n
\n

\u6765\u6e90\uff1a\u96c6\u667a\u767e\u79d1<\/p>\n

\u7f16\u8f91\uff1a\u738b\u5efa\u840d<\/p>\n<\/section>\n<\/blockquote>\n<\/section>\n


<\/section>\n

<\/section>\n
\u63a8\u8350\u9605\u8bfb<\/span><\/strong><\/span><\/section>\n