{"id":37602,"date":"2022-09-30T19:05:13","date_gmt":"2022-09-30T11:05:13","guid":{"rendered":"https:\/\/swarma.org\/?p=37602"},"modified":"2022-09-30T19:05:13","modified_gmt":"2022-09-30T11:05:13","slug":"pcic2022%e7%ac%ac%e5%9b%9b%e5%b1%8a%e6%b3%9b%e5%a4%aa%e5%b9%b3%e6%b4%8b%e5%9b%a0%e6%9e%9c%e6%8e%a8%e6%96%ad%e5%a4%a7%e4%bc%9a%e6%80%bb%e7%bb%93","status":"publish","type":"post","link":"https:\/\/swarma.org\/?p=37602","title":{"rendered":"PCIC2022\u7b2c\u56db\u5c4a\u6cdb\u592a\u5e73\u6d0b\u56e0\u679c\u63a8\u65ad\u5927\u4f1a\u603b\u7ed3"},"content":{"rendered":"
\n
\n
\n
\n

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

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

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

\u5bfc\u8bed<\/strong><\/p>\n<\/section>\n<\/section>\n

\n

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

\u4e3a\u4e86\u66f4\u591a\u5730\u63a8\u52a8\u56e0\u679c\u63a8\u65ad\u5b66\u79d1\u7684\u53d1\u5c55\uff0c\u805a\u96c6\u56fd\u5185\u5916\u56e0\u679c\u63a8\u65ad\u7684\u4e00\u7ebf\u79d1\u7814\u5de5\u4f5c\u8005\uff0c\u5171\u540c\u8ba8\u8bba\u56e0\u679c\u79d1\u5b66\u7684\u6700\u65b0\u8fdb\u5c55\uff0c\u5317\u4eac\u5927\u5b66\u8bb2\u5e2d\u6559\u6388\u3001\u5317\u4eac\u5927\u5b66\u516c\u5171\u536b\u751f\u5b66\u9662\u751f\u7269\u7edf\u8ba1\u7cfb\u4e3b\u4efb\u3001\u5317\u4eac\u5927\u5b66\u5317\u4eac\u56fd\u9645\u6570\u5b66\u7814\u7a76\u4e2d\u5fc3\u751f\u7269\u7edf\u8ba1\u548c\u4fe1\u606f\u7814\u7a76\u5ba4\u4e3b\u4efb\u5468\u6653\u534e\u7b49\u53d1\u8d77\u4e86\u6cdb\u592a\u5e73\u6d0b\u56e0\u679c\u63a8\u65ad\u5927\u4f1a<\/a>\uff0c\u4f1a\u8bae\u4e8e2021\u5e749\u670817\u65e5\u81f318\u65e5\u4e3e\u884c\u3002\u6cdb\u592a\u5e73\u6d0b\u56e0\u679c\u63a8\u65ad\u5927\u4f1a\u662f\u56e0\u679c\u63a8\u65ad\u9886\u57df\u56fd\u9645\u4e0a\u6700\u9886\u5148\u7684\u4f1a\u8bae\u4e4b\u4e00\u3002\u672c\u6b21\u4f1a\u8bae\u5171\u9080\u8bf7\u4e8637\u4f4d\u5609\u5bbe\u4f5c\u62a5\u544a\uff0c\u63a2\u8ba8\u56e0\u679c\u63a8\u65ad\u9886\u57df\u7684\u65b0\u8fdb\u5c55\u3002\u4f1a\u8bae\u4e0d\u4ec5\u8bbe\u6709\u4e13\u5bb6\u5b66\u8005\u62a5\u544a\uff0c\u8fd8\u53d1\u5e03\u4e86\u56e0\u679c\u63a8\u7406\u7ade\u8d5b\u8d5b\u9898\uff08PCIC 2022\u534e\u4e3a\u56e0\u679c\u63a8\u7406\u6311\u6218\u8d5b<\/a>\uff09\u3002\u7d2f\u8ba1\u8d85\u8fc71000\u4eba\u6b21\u6ce8\u518c\u4f1a\u8bae\uff0c\u8d85\u8fc73\u4e07\u4eba\u6b21\u89c2\u770b\u4f1a\u8bae\u76f4\u64ad\u3002<\/span><\/p>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n

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


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


<\/section>\n
\u4ee5\u4e0b\u6211\u4eec\u4e3a\u5927\u5bb6\u68b3\u7406\u4e86\u672c\u6b21\u56e0\u679c\u63a8\u65ad\u5927\u4f1a\u7684\u62a5\u544a\u6458\u8981\u4e0e\u53c2\u8003\u8d44\u6599\u3002\u7ecf\u8fc7\u62a5\u544a\u4eba\u7684\u540c\u610f\uff0c\u540e\u7eed\u6211\u4eec\u4f1a\u5c06\u90e8\u5206\u56de\u653e\u4e0ePPT\u8d44\u6599\u516c\u5f00\u5728PCIC\u5927\u4f1a\u7684\u5b98\u65b9\u7f51\u7ad9\uff08https:\/\/pattern.swarma.org\/pcic\/\uff09\uff0c\u656c\u8bf7\u671f\u5f85\u3002<\/span><\/section>\n
<\/section>\n
\u56e0\u679c\u63a8\u65ad\u5927\u4f1a\u5b98\u65b9\u7f51\u7ad9<\/span><\/section>\n

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


<\/p>\n


<\/section>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u4e00\uff08Linbo Wang\uff0cChair\uff09<\/strong><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Estimating the Optimal Dynamic Treatment Regime with Restrictions using Observational Data<\/span><\/strong><\/span><\/h4>\n
Lu Wang<\/span><\/section>\n
\u6700\u4f18\u52a8\u6001\u6cbb\u7597\u65b9\u6848\u662f\u5728\u75be\u75c5\u6cbb\u7597\u8fc7\u7a0b\u4e2d\u975e\u5e38\u91cd\u8981\u7684\u95ee\u9898\u3002\u5728\u6cbb\u7597\u8fc7\u7a0b\u4e2d\uff0c\u533b\u751f\u9700\u8981\u6839\u636e\u6cbb\u7597\u8fc7\u7a0b\u4e2d\u641c\u96c6\u5230\u7684\u75c5\u4eba\u4fe1\u606f\u6765\u51b3\u5b9a\u4e0b\u4e00\u6b65\u7684\u6cbb\u7597\u65b9\u6cd5\uff0c\u4ece\u800c\u8fbe\u5230\u6700\u597d\u7684\u6cbb\u7597\u6548\u679c\u3002\u738b\u6559\u6388\u5728\u8be5\u65b9\u5411\u63d0\u51fa\u4e86\u4e00\u7cfb\u5217\u65b9\u6cd5\u6765\u89e3\u51b3\u6b64\u7c7b\u95ee\u9898\uff0c\u6bd4\u5982\u57fa\u4e8e\u51b3\u7b56\u6811\u7684\u5f3a\u5316\u5b66\u4e60\u65b9\u6cd5\uff08T-RL\uff09\uff0c\u8be5\u65b9\u6cd5\u53ef\u4ee5\u5904\u7406\u591a\u503c\u5904\u7406\u53d8\u91cf\uff0c\u4e14\u901a\u8fc7\u7ed3\u5408\u534a\u53c2\u56de\u5f52\u4f30\u8ba1\u91cf\u548c\u975e\u53c2\u673a\u5668\u5b66\u4e60\u65b9\u6cd5\u8fbe\u5230\u7a33\u5065\u6709\u6548\u4f30\u8ba1\u3002\u968f\u540e\u738b\u6559\u6388\u53c8\u63d0\u51fa\u4e86\u968f\u673a\u51b3\u7b56\u6811\u5f3a\u5316\u5b66\u4e60\u65b9\u6cd5\uff08ST-RL\uff09\uff0c\u8be5\u65b9\u6cd5\u5229\u7528MCMC\u65b9\u6cd5\uff0c\u4e0d\u6613\u8fc7\u62df\u5408\u3002\u4f46\u662f\u5728\u5b9e\u9645\u5e94\u7528\u8fc7\u7a0b\u4e2d\uff0c\u5f80\u5f80\u4f1a\u78b0\u5230\u4e00\u4e9b\u9650\u5236\uff0c\u4f8b\u5982\u67d0\u4e9b\u53d8\u91cf\u4e0d\u5e94\u8be5\u7528\u6765\u4f5c\u4e3a\u51b3\u7b56\u4f9d\u636e\uff0c\u6216\u8005\u67d0\u4e9b\u6cbb\u7597\u65b9\u6cd5\u5bf9\u4e8e\u5c06\u6765\u7684\u75c5\u4eba\u5e76\u4e0d\u9002\u7528\u7b49\u3002\u6b64\u65f6\u4f20\u7edf\u7684\u65b9\u6cd5\u5e76\u4e0d\u80fd\u89e3\u51b3\u6b64\u7c7b\u95ee\u9898\u3002\u56e0\u6b64\u738b\u8001\u5e08\u63d0\u51fa\u4e86\u65b0\u7684\u5f3a\u5316\u5b66\u4e60\u65b9\u6cd5\uff08RT-RL\uff09\uff0c\u8be5\u65b9\u6cd5\u5728\u4f18\u5316\u8fc7\u7a0b\u4e2d\u5c06\u8fd9\u4e9b\u9650\u5236\u6761\u4ef6\u8003\u8651\u5728\u5176\u4e2d\uff0c\u4ece\u800c\u80fd\u591f\u89e3\u51b3\u5b9e\u9645\u5e94\u7528\u8fc7\u7a0b\u4e2d\u7684\u95ee\u9898\u3002\u8be5\u65b9\u6cd5\u5728\u5b9e\u9645\u6570\u636e\u5206\u6790\u4e2d\u53d6\u5f97\u4e86\u5f88\u597d\u7684\u6548\u679c\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Nina Zhou, Lu Wang, Daniel Almirall (2022) Estimating tree-based dynamic treatment regimes using observational data with restricted treatment sequences Biometrics<\/span><\/section>\n


<\/h4>\n

Fighting Noise with Noise: Causal Inference with Many Candidate Instruments<\/span><\/strong><\/span><\/h4>\n
Dehan Kong<\/span><\/section>\n
\u5728\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u7814\u7a76\u4e2d\uff0c\u4f20\u7edf\u505a\u6cd5\u662f\u5148\u901a\u8fc7\u76f8\u5173\u6027\u7684\u5f3a\u5f31\u7b5b\u9009\u51fa\u53ef\u80fd\u7684\u5de5\u5177\u53d8\u91cf\uff0c\u7136\u540e\u5728\u901a\u8fc72SLS\u65b9\u6cd5\u627e\u5230\u591a\u4e2a\u56e0\u679c\u4f5c\u7528\u7684\u4f30\u8ba1\uff0c\u6700\u540e\u4ee5\u5176\u4e2d\u7684\u4f17\u6570\u4f5c\u4e3a\u4f30\u8ba1\u91cf\u3002\u7136\u800c\u5728\u5b9e\u9645\u5206\u6790\u4e2d\uff0c\u7531\u4e8e\u7b2c\u4e00\u6b65\u7684\u7b5b\u9009\u8fc7\u7a0b\uff0c\u4f7f\u5f97\u67d0\u4e9b\u65e0\u5173\u5de5\u5177\u53d8\u91cf\u4e0e\u5904\u7406\u53d8\u91cf\u6709\u53ef\u7591\u7684\u76f8\u5173\u6027\uff0c\u6700\u7ec8\u5bfc\u81f4\u4f30\u8ba1\u91cf\u6709\u504f\u3002\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\uff0c\u5b54\u6559\u6388\u63d0\u51fa\u5229\u7528\u901a\u8fc7\u751f\u6210\u4f2a\u53d8\u91cf\u6765\u6d88\u9664\u65e0\u5173\u5de5\u5177\u53d8\u91cf\u9020\u6210\u7684\u504f\u5dee\uff0c\u5e76\u8bc1\u660e\u4e86\u8be5\u65b9\u6cd5\u7684\u7406\u8bba\u6027\u8d28\u3002\u5b54\u6559\u6388\u7528\u8be5\u65b9\u6cd5\u5206\u6790\u4e86BMI\u5bf9\u4e8e\u751f\u6d3b\u8d28\u91cf\u7684\u5f71\u54cd\u7814\u7a76\u4e2d\uff0c\u5f97\u5230\u4e86\u6bd4\u4e4b\u524d\u65b9\u6cd5\u66f4\u597d\u7684\u7ed3\u679c\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Xinyi Zhang, Linbo Wang, Stanislav Volgushev and Dehan Kong. (2022+). Fighting Noise with Noise: Causal Inference with Many Candidate Instruments.<\/span><\/section>\n


<\/h4>\n

Outcome-adaptive LASSO for confounder selection with time-varying treatments<\/span><\/strong><\/span><\/h4>\n
Mireille Schnitzer<\/span><\/section>\n
\u968f\u65f6\u95f4\u53d8\u5316\u7684\u5904\u7406\u5728\u751f\u7269\u7edf\u8ba1\u9886\u57df\u662f\u5341\u5206\u91cd\u8981\u7684\u7814\u7a76\u65b9\u5411\u3002\u6559\u6388\u5229\u7528\u9006\u6982\u7387\u52a0\u6743\u65b9\u6cd5\u6765\u4f30\u8ba1\u5173\u5fc3\u7684\u53c2\u6570\u3002\u7531\u4e8e\u6a21\u578b\u4e2d\u7684\u534f\u53d8\u91cf\u4f1a\u968f\u65f6\u95f4\u589e\u52a0\uff0c\u5728\u4f30\u8ba1\u8fc7\u7a0b\u4e2d\uff0c\u5982\u4f55\u9009\u62e9\u6df7\u6742\u53d8\u91cf\u4f7f\u5f97\u4f30\u8ba1\u91cf\u7684\u65b9\u5dee\u51cf\u5c0f\u540c\u65f6\u4e0d\u589e\u52a0\u504f\u5dee\u662f\u5341\u5206\u91cd\u8981\u7684\u95ee\u9898\u3002\u6559\u6388\u63d0\u51fa\u5229\u7528adaptive lasso\u65b9\u6cd5\u6765\u8fdb\u884c\u53d8\u91cf\u9009\u62e9\u3002\u4e3a\u4e86\u5728\u5b9e\u9645\u5206\u6790\u4e2d\u80fd\u591f\u4f7f\u7528\u8be5\u65b9\u6cd5\uff0c\u6559\u6388\u63d0\u51fa\u4e86\u9009\u62e9adaptive lasso\u65b9\u6cd5\u4e2d\u8d85\u53c2\u6570\u7684\u9009\u53d6\u65b9\u6cd5\u3002\u8be5\u65b9\u6cd5\u5728\u6a21\u62df\u548c\u5b9e\u9645\u6570\u636e\u5206\u6790\u4e2d\u53d6\u5f97\u4e86\u826f\u597d\u7684\u8868\u73b0\u3002\u540c\u65f6\u6559\u6388\u5728\u6700\u540e\u4e5f\u63d0\u51fa\u4e86\u76ee\u524d\u7684\u65b9\u6cd5\u7684\u9650\u5236\uff0c\u4f8b\u5982\u4f9d\u8d56\u4e8e\u53c2\u6570\u6a21\u578b\uff0c\u53c2\u6570\u4f30\u8ba1\u901f\u7387\u7b49\u3002<\/span><\/section>\n

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

Estimating heterogeneous treatment effects with right-censored data via causal survival forests<\/span><\/strong><\/span><\/h4>\n
Yifan Cui<\/span><\/section>\n
\u5f02\u8d28\u6027\u56e0\u679c\u6548\u5e94\u5728\u7cbe\u51c6\u533b\u7597\uff0c\u7ecf\u6d4e\uff0c\u5fc3\u7406\uff0c\u79d1\u6280\u5e73\u53f0\u7b49\u9886\u57df\u90fd\u6709\u7740\u5e7f\u6cdb\u7684\u5e94\u7528\u3002\u5d14\u6559\u6388\u4ecb\u7ecd\u4e86\u6700\u8fd1\u5c06\u56e0\u679c\u68ee\u6797\u5e94\u7528\u4e8e\u751f\u5b58\u5206\u6790\u7684\u5de5\u4f5c\u3002\u901a\u8fc7\u5047\u5b9a\u5220\u5931\u65f6\u95f4\u4e0e\u751f\u5b58\u65f6\u95f4\u6761\u4ef6\u72ec\u7acb\uff0c\u5f02\u8d28\u6027\u751f\u5b58\u65f6\u95f4\u5dee\u5f02\u80fd\u591f\u5728\u53ef\u5ffd\u7565\u6027\u7b49\u5047\u5b9a\u4e0b\u5f97\u5230\u8bc6\u522b\u3002\u5728\u5f97\u5230\u53c2\u6570\u7684\u8bc6\u522b\u6027\u540e\uff0c\u5d14\u6559\u6388\u63d0\u51fa\u5229\u7528\u9006\u6982\u7387\u52a0\u6743\u56e0\u679c\u68ee\u6797\u65b9\u6cd5\uff0c\u4ee5\u53ca\u589e\u5e7f\u9006\u6982\u7387\u52a0\u6743\u65b9\u6cd5\uff08augmented IPW\uff09\u3002\u6700\u540e\u5c06\u63d0\u51fa\u7684\u65b9\u6cd5\u4f7f\u7528\u5230\u4e86\u5173\u4e8e\u9ad8\u8840\u538b\u7684\u5b9e\u9645\u6570\u636e\u5206\u6790\u4e2d\u3002<\/span><\/section>\n


<\/h4>\n

The synthetic instrument<\/span><\/strong><\/span><\/h4>\n
Linbo Wang<\/span><\/section>\n
\u591a\u5904\u7406\u5b58\u5728\u65f6\u7684\u56e0\u679c\u63a8\u65ad\u5728\u6700\u8fd1\u5f15\u8d77\u4e86\u5f88\u591a\u5173\u6ce8\u3002\u5f53\u5b58\u5728\u672a\u89c2\u6d4b\u6df7\u6742\u65f6\uff0c \u738b\u6559\u6388\u8003\u8651\u4e86\u7ebf\u6027\u6a21\u578b\u4e0b\u56e0\u679c\u4f5c\u7528\u7684\u4f30\u8ba1\u95ee\u9898\u3002\u5728\u591a\u5904\u7406\u6ee1\u8db3\u56e0\u5b50\u6a21\u578b\uff0c\u4e14\u591a\u4e2a\u5904\u7406\u4e2d\u6709\u90e8\u5206\u5904\u7406\u5bf9\u4e8e\u7ed3\u5c40\u6ca1\u6709\u56e0\u679c\u4f5c\u7528\u65f6\uff0c\u738b\u6559\u6388\u63d0\u51fa\u4e86\u5408\u6210\u5de5\u5177\u53d8\u91cf\u7684\u4f30\u8ba1\u65b9\u6cd5\u3002\u8be5\u65b9\u6cd5\u53ef\u4ee5\u89e3\u51b3\u9ad8\u7ef4\u5904\u7406\u5b58\u5728\u65f6\u7684\u4f30\u8ba1\u95ee\u9898\uff0c\u5e76\u5728\u6a21\u62df\u4e2d\u5177\u6709\u826f\u597d\u7684\u8868\u73b0\u3002<\/span><\/section>\n


<\/h2>\n

<\/section>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u4e8c\uff08Jinzhu Jia\uff0cChair\uff09<\/strong><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

No Star Is Good News: A Unified Look at Rerandomization Based on P-Values from Covariate Balance Tests<\/span><\/strong><\/span><\/h4>\n
Anqi Zhao, National University of Singapore<\/span><\/section>\n
\u968f\u673a\u5bf9\u7167\u8bd5\u9a8c\u5728\u5e73\u5747\u610f\u4e49\u4e0b\u5e73\u8861\u6240\u6709\u534f\u53d8\u91cf\uff0c\u4e3a\u8bc4\u4f30\u6cbb\u7597\u6548\u679c\u63d0\u4f9b\u4e86\u91d1\u6807\u51c6\u3002\u7136\u800c\uff0c\u673a\u4f1a\u4e0d\u5e73\u8861\u6216\u591a\u6216\u5c11\u5730\u5b58\u5728\u4e8e\u5df2\u5b9e\u73b0\u7684\u6cbb\u7597\u5206\u914d\u4e2d\uff0c\u4f7f\u968f\u540e\u7684\u63a8\u65ad\u53ef\u80fd\u5177\u6709\u8f83\u5927\u7684\u6ce2\u52a8\u6027\u3002\u73b0\u4ee3\u79d1\u5b66\u51fa\u7248\u7269\u8981\u6c42\u62a5\u544a\u534f\u53d8\u91cf\u5e73\u8861\u8868\uff0c\u4e0d\u4ec5\u5305\u62ec\u6cbb\u7597\u7ec4\u548c\u5bf9\u7167\u7ec4\u7684\u534f\u53d8\u91cf\u5747\u503c\uff0c\u8fd8\u5305\u62ec\u4e24\u7ec4\u4e4b\u95f4\u5176\u5dee\u5f02\u663e\u8457\u6027\u68c0\u9a8c\u7684\u76f8\u5173p\u503c\u3002\u5728\u5b9e\u9645\u9700\u6c42\u4e2d\uff0c\u6211\u4eec\u9700\u8981\u907f\u514d\u5c0fp\u503c\u5e26\u6765\u7684\u534f\u53d8\u91cf\u5728\u5bf9\u7167\u7ec4\u548c\u6cbb\u7597\u7ec4\u4e4b\u95f4\u4e0d\u5e73\u8861\u7684\u73b0\u8c61\uff0c\u4ece\u800c\u51cf\u5c11\u540e\u7eed\u5bf9\u6cbb\u7597\u6548\u679c\u4f30\u8ba1\u7684\u504f\u5dee\u3002\u907f\u514d\u5c0fp\u503c\u7684\u5b9e\u9645\u9700\u6c42\u4f7f\u5f97\u901a\u8fc7\u5047\u8bbe\u68c0\u9a8c\u8fdb\u884c\u534f\u53d8\u91cf\u7684\u5e73\u8861\u68c0\u67e5\u548c\u518d\u968f\u673a\u5206\u6790\u6210\u4e3a\u6539\u5584\u968f\u673a\u5bf9\u7167\u8bd5\u9a8c\u4e2d\u534f\u53d8\u91cf\u5e73\u8861\u7684\u4e00\u4e2a\u6709\u5438\u5f15\u529b\u7684\u5de5\u5177\u3002\u6211\u4eec\u7814\u7a76\u4e86\u57fa\u4e8e\u534f\u53d8\u91cf\u518d\u5e73\u8861\u6d4b\u8bd5\u7684p\u503c\uff08Rerandomization based on p-values, ReP\uff09\u7684\u5404\u79cd\u53ef\u80fd\u6709\u7528\u7684\u518d\u968f\u673a\u65b9\u6848\uff0c\u4e0e\u4e0d\u8fdb\u884c\u518d\u5e73\u8861\u7684\u65b9\u6848\u8fdb\u884c\u6bd4\u8f83\uff0c\u5e76\u8bc1\u660e\u4e86\u5b83\u4eec\u5bf9\u540e\u7eed\u7edf\u8ba1\u63a8\u65ad\u7684\u5f71\u54cd\uff0c\u5176\u4e2d\u4e3b\u8981\u8fdb\u884c\u6bd4\u8f83\u7684\u4e09\u79cd\u4f30\u8ba1\u91cf\u5206\u522b\u662f\u53ea\u8003\u8651\u5206\u7ec4\u4fe1\u606f\u7684Neyman\u4f30\u8ba1\u91cf\u3001\u8003\u8651\u534f\u53d8\u91cf\u4fe1\u606f\u7684Fisher\u4f30\u8ba1\u91cf\u3001\u4ee5\u53ca\u8003\u8651\u4ea4\u4e92\u4f5c\u7528\u7684Lin\u4f30\u8ba1\u91cf\u3002\u6211\u4eec\u7684\u4e3b\u8981\u53d1\u73b0\u6709\u4e24\u65b9\u9762\u3002\u9996\u5148\uff0c\u5728\u6240\u6709\u68c0\u67e5\u7684ReP\u65b9\u6848\u4e0b\uff0c\u8003\u8651\u5b8c\u5168\u76f8\u4e92\u4f5c\u7528\u56de\u5f52\u7684Lin\u4f30\u8ba1\u91cf\u662f\u6e10\u8fd1\u6700\u6709\u6548\u7684\uff0c\u5e76\u4e14\u5141\u8bb8\u65b9\u4fbf\u7684\u56de\u5f52\u8f85\u52a9\uff0cLin\u4f30\u8ba1\u91cf\u7684\u7edf\u8ba1\u63a8\u65ad\u7ed3\u679c\u5728ReP\u6846\u67b6\u4e0b\u548c\u5728\u5b8c\u5168\u7684\u968f\u673a\u5316\u6846\u67b6\u4e0b\u662f\u4e00\u81f4\u7684\u3002\u5176\u6b21\uff0cReP\u4e0d\u4ec5\u63d0\u9ad8\u4e86\u534f\u53d8\u91cf\u5e73\u8861\uff0c\u8fd8\u63d0\u9ad8\u4e86Neyman\u4f30\u8ba1\u91cf\u548cFisher\u4f30\u8ba1\u91cf\u7684\u6548\u7387\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Zhao, A., & Ding, P. (2021). No star is good news: A unified look at rerandomization based on p-values from covariate balance tests. arXiv preprint arXiv:2112.10545.<\/span><\/section>\n


<\/h4>\n

Design-based theory for cluster rerandomization<\/span><\/strong><\/span><\/h4>\n
Hanzhong Liu, Tsinghua University<\/span><\/section>\n
\u5b8c\u5168\u968f\u673a\u5316\u5e73\u5747\u5e73\u8861\u534f\u53d8\u91cf\uff0c\u4f46\u5728\u6709\u9650\u6837\u672c\u4e2d\u5f80\u5f80\u5b58\u5728\u534f\u53d8\u91cf\u4e0d\u5e73\u8861\u3002\u518d\u968f\u673a\u5316\u53ef\u4ee5\u901a\u8fc7\u5220\u9664\u90e8\u5206\u6cbb\u7597\u5206\u914d\u65b9\u6848\u6765\u786e\u4fdd\u5df2\u5b9e\u73b0\u5b9e\u9a8c\u4e2d\u7684\u534f\u53d8\u91cf\u5e73\u8861\u3002\u7531\u4e8e\u540e\u52e4\u9650\u5236\u6216\u653f\u7b56\u8003\u8651\uff0c\u516c\u5171\u536b\u751f\u548c\u793e\u4f1a\u79d1\u5b66\u9886\u57df\u7684\u8bb8\u591a\u5b9e\u5730\u5b9e\u9a8c\u5c06\u6cbb\u7597\u5b89\u6392\u5728\u96c6\u7fa4\u4e00\u7ea7\u3002\u6b64\u5916\uff0c\u5b83\u4eec\u7ecf\u5e38\u5728\u8bbe\u8ba1\u9636\u6bb5\u4e0e\u518d\u968f\u673a\u5b9e\u9a8c\u7ec4\u5408\u3002\u6211\u4eec\u5c06\u805a\u7c7b\u518d\u968f\u673a\u5316\u5b9a\u4e49\u4e3a\u4e00\u4e2a\u4e0e\u518d\u968f\u673a\u5316\u76f8\u7ed3\u5408\u7684\u805a\u7c7b\u968f\u673a\u5b9e\u9a8c\uff0c\u4ee5\u5e73\u8861\u4e2a\u4f53\u6216\u805a\u7c7b\u6c34\u5e73\u4e0a\u7684\u534f\u53d8\u91cf\u3002\u73b0\u6709\u7684\u6e10\u8fd1\u7406\u8bba\u53ea\u80fd\u5904\u7406\u5728\u4e2a\u4f53\u6c34\u5e73\u4e0a\u5206\u914d\u7684\u5904\u7406\u7684\u518d\u968f\u673a\u5316\uff0c\u8fd9\u4f7f\u5f97\u96c6\u7fa4\u518d\u968f\u673a\u5316\u6210\u4e3a\u4e00\u4e2a\u5f00\u653e\u95ee\u9898\u3002\u4e3a\u4e86\u586b\u8865\u8fd9\u4e00\u7a7a\u767d\uff0c\u6211\u4eec\u63d0\u4f9b\u4e86\u4e00\u79cd\u57fa\u4e8e\u8bbe\u8ba1\u7684\u7c07\u91cd\u968f\u673a\u7406\u8bba\u3002\u6b64\u5916\uff0c\u6211\u4eec\u6bd4\u8f83\u4e86\u4e24\u79cd\u4f7f\u7528\u5173\u4e8e\u534f\u53d8\u91cf\u91cd\u8981\u6027\u7684\u5148\u9a8c\u4fe1\u606f\u7684\u805a\u7c7b\u518d\u968f\u673a\u65b9\u6848\uff1a\u4e00\u79cd\u57fa\u4e8e\u52a0\u6743\u6b27\u51e0\u91cc\u5fb7\u8ddd\u79bb\uff0c\u53e6\u4e00\u79cd\u57fa\u4e8e\u5177\u6709\u534f\u53d8\u91cf\u5c42\u7684\u9a6c\u6c0f\u8ddd\u79bb\u3002\u6211\u4eec\u8bc1\u660e\u4e86\u524d\u8005\u5728\u6700\u4f18\u6743\u91cd\u548c\u6b63\u4ea4\u5316\u534f\u53d8\u91cf\u60c5\u5f62\u4e0b\u663e\u8457\u4f18\u4e8e\u540e\u8005\u3002\u6700\u540e\uff0c\u6211\u4eec\u8ba8\u8bba\u4e86\u534f\u53d8\u91cf\u8c03\u6574\u5728\u5206\u6790\u9636\u6bb5\u7684\u4f5c\u7528\uff0c\u5e76\u63a8\u8350\u4e86\u53ef\u901a\u8fc7\u6700\u5c0f\u4e8c\u4e58\u6cd5\uff08\u5229\u7528\u76f8\u5173\u7684\u7a33\u5065\u6807\u51c6\u8bef\u5dee\uff09\u6765\u65b9\u4fbf\u5b9e\u65bd\u7684\u534f\u53d8\u91cf\u8c03\u6574\u7a0b\u5e8f\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Lu, X., Liu, T., Liu, H., & Ding, P. Design-based theory for cluster rerandomization. Biometrika. Forthcoming.<\/span><\/section>\n


<\/h4>\n

Identification of Linear Latent Hierarchical Structure<\/span><\/strong><\/span><\/h4>\n
Feng Xie, Beijing Technology and Business University<\/span><\/section>\n
\u4f20\u7edf\u7684\u56e0\u679c\u53d1\u73b0\u65b9\u6cd5\u4e3b\u8981\u4fa7\u91cd\u4e8e\u4f30\u8ba1\u6d4b\u91cf\u53d8\u91cf\u4e4b\u95f4\u7684\u56e0\u679c\u5173\u7cfb\uff0c\u4f46\u5728\u8bb8\u591a\u73b0\u5b9e\u95ee\u9898\u4e2d\uff0c\u5982\u57fa\u4e8e\u95ee\u5377\u7684\u5fc3\u7406\u6d4b\u91cf\u7814\u7a76\uff0c\u6d4b\u91cf\u53d8\u91cf\u662f\u7531\u56e0\u679c\u76f8\u5173\u7684\u6f5c\u5728\u53d8\u91cf\u751f\u6210\u7684\u3002\u5728\u8fd9\u7bc7\u6f14\u8bb2\u4e2d\uff0c\u6211\u4eec\u5c06\u7814\u7a76\u53d1\u73b0\u9690\u85cf\u7684\u56e0\u679c\u53d8\u91cf\u548c\u4f30\u8ba1\u56e0\u679c\u7ed3\u6784\u7684\u95ee\u9898\uff0c\u5305\u62ec\u6f5c\u5728\u53d8\u91cf\u4e4b\u95f4\u7684\u56e0\u679c\u5173\u7cfb\u4ee5\u53ca\u6f5c\u5728\u53d8\u91cf\u548c\u6d4b\u91cf\u53d8\u91cf\u4e4b\u95f4\u7684\u5173\u7cfb\u3002\u6211\u4eec\u653e\u677e\u4e86\u5e38\u7528\u7684\u6d4b\u91cf\u5047\u8bbe\uff0c\u5141\u8bb8\u6f5c\u5728\u53d8\u91cf\u7684\u5b50\u53d8\u91cf\u4e5f\u662f\u6f5c\u5728\u7684\uff0c\u56e0\u6b64\u5904\u7406\u4e00\u79cd\u7279\u5b9a\u7c7b\u578b\u7684\u6f5c\u5728\u5c42\u6b21\u56e0\u679c\u7ed3\u6784\u3002\u7279\u522b\u5730\uff0c\u6211\u4eec\u5b9a\u4e49\u4e86\u4e00\u4e2a\u6700\u5c0f\u6f5c\u5728\u5c42\u6b21\u7ed3\u6784\uff0c\u5e76\u8bc1\u660e\u4e86\u5bf9\u4e8e\u5177\u6709\u6700\u5c0f\u6f5c\u5728\u5c42\u6b21\u7684\u7ebf\u6027\u975e\u9ad8\u65af\u6a21\u578b\uff0c\u6574\u4e2a\u7ed3\u6784\u4ec5\u53ef\u4ece\u6d4b\u91cf\u53d8\u91cf\u8bc6\u522b\u3002\u6b64\u5916\uff0c\u6211\u4eec\u5f00\u53d1\u4e86\u4e00\u79cd\u539f\u7406\u65b9\u6cd5\uff0c\u901a\u8fc7\u4ee5\u7279\u5b9a\u65b9\u5f0f\u6d4b\u8bd5\u5e7f\u4e49\u72ec\u7acb\u566a\u58f0\uff08Generalized Independent Noise, GIN\uff09\u6761\u4ef6\u6765\u8bc6\u522b\u7ed3\u6784\u3002\u901a\u8fc7\u5728\u5408\u6210\u6570\u636e\u548c\u771f\u5b9e\u6570\u636e\u8fdb\u884c\u5145\u5206\u7684\u5b9e\u9a8c\uff0c\u6211\u4eec\u8bc1\u660e\u4e86\u6240\u63d0\u51fa\u7684\u65b9\u6cd5\u7684\u6709\u6548\u6027\u3002\u6211\u4eec\u7684\u65b9\u6cd5\u4e0d\u4ec5\u4ec5\u4e3a\u7ed3\u6784\u53ef\u8bc6\u522b\u6027\u63d0\u4f9b\u5145\u5206\u6761\u4ef6\uff0c\u540c\u65f6\u57fa\u4e8e\u7406\u8bba\u63a8\u5bfc\u548c\u5b9e\u9a8c\u8bba\u8bc1\u4e86\u5b66\u4e60\u7ebf\u6027\u6f5c\u5728\u5c42\u6b21\u7ed3\u6784\u7684\u91cd\u8981\u6027\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Xie, F., Huang, B., Chen, Z., He, Y., Geng, Z., & Zhang, K. (2022). Identification of Linear Non-Gaussian Latent Hierarchical Structure. In International Conference on Machine Learning (pp. 24370-24387). PMLR.<\/span><\/section>\n


<\/h4>\n

Paradoxes and resolutions for semiparametric data fusion with individual data and summary statistics<\/span><\/strong><\/span><\/h4>\n
Wang Miao, Peking University<\/span><\/section>\n
\u5047\u8bbe\u6211\u4eec\u6709\u6765\u81ea\u5185\u90e8\u7814\u7a76\u7684\u4e2a\u4f53\u6570\u636e\u548c\u6765\u81ea\u76f8\u5173\u5916\u90e8\u7814\u7a76\u7684\u5404\u7c7b\u6c47\u603b\u7edf\u8ba1\u6570\u636e\u3002\u5916\u90e8\u6c47\u603b\u7edf\u8ba1\u91cf\u88ab\u7528\u4f5c\u5185\u90e8\u6570\u636e\u5206\u5e03\u76845\u4e2a\u7ea6\u675f\u6761\u4ef6\uff0c\u8fd9\u6709\u52a9\u4e8e\u6539\u8fdb\u7edf\u8ba1\u63a8\u65ad\uff1b\u7136\u800c\uff0c\u8fd9\u79cd\u6570\u636e\u6574\u5408\u4f1a\u4ea7\u751f\u77db\u76fe\u7684\u7ed3\u679c\uff1a\u5982\u679c\u6c47\u603b\u7edf\u8ba1\u6570\u636e\u7684\u4e0d\u786e\u5b9a\u6027\u4e0d\u53ef\u5ffd\u7565\uff0c\u5219\u53ef\u80fd\u4f1a\u53d1\u751f\u6548\u7387\u635f\u5931\uff0c\u5982\u679c\u4ece\u5185\u90e8\u7814\u7a76\u7684\u4e0d\u540c\u4eba\u7fa4\u4e2d\u83b7\u5f97\uff0c\u5219\u53ef\u80fd\u51fa\u73b0\u4f30\u8ba1\u504f\u5dee\u3002\u6211\u4eec\u5728\u534a\u53c2\u6570\u6846\u67b6\u4e2d\u7814\u7a76\u4e86\u8fd9\u4e9b\u77db\u76fe\u7684\u7ed3\u679c\u3002\u6211\u4eec\u5efa\u7acb\u4e86\u4f30\u8ba1\u5185\u90e8\u6570\u636e\u5206\u5e03\u7684\u4e00\u822c\u51fd\u6570\u7684\u534a\u53c2\u6570\u6548\u7387\u754c\uff0c\u8bc1\u660e\u4e86\u5176\u4e0d\u5927\u4e8e\u4ec5\u4f7f\u7528\u5185\u90e8\u6570\u636e\u7684\u6548\u7387\u754c\u3002\u6211\u4eec\u63d0\u51fa\u4e86\u4e00\u79cd\u6570\u636e\u878d\u5408\u7684\u6709\u6548\u4f30\u8ba1\u91cf\uff0c\u8be5\u4f30\u8ba1\u91cf\u8fbe\u5230\u4e86\u8fd9\u4e2a\u754c\u9650\uff0c\u4ece\u800c\u89e3\u51b3\u4e86\u6548\u7387\u6096\u8bba\u3002\u8be5\u521d\u59cb\u6570\u636e\u878d\u5408\u7684\u6709\u6548\u4f30\u8ba1\u91cf\u901a\u8fc7\u81ea\u9002\u5e94lasso\u60e9\u7f5a\u8fdb\u4e00\u6b65\u6b63\u5219\u5316\uff0c\u4ece\u800c\u5bfc\u81f4\u7684\u4f30\u8ba1\u5668\u53ef\u4ee5\u5b9e\u73b0\u4e0e\u4ec5\u4f7f\u7528\u65e0\u504f\u6c47\u603b\u7edf\u8ba1\u7684oracle\u4f30\u8ba1\u5668\u76f8\u540c\u7684\u6e10\u8fd1\u5206\u5e03\uff0c\u8fd9\u89e3\u51b3\u4e86\u504f\u5dee\u6096\u8bba\u3002\u6211\u4eec\u5206\u522b\u5728\u6a21\u62df\u6570\u636e\u548c\u5bf9\u5e7d\u95e8\u87ba\u6746\u83cc\u611f\u67d3\u6570\u636e\u4e2d\u5e94\u7528\u4e86\u6211\u4eec\u7684\u65b9\u6cd5\uff0c\u5176\u7ed3\u679c\u8bc1\u660e\u4e86\u6240\u63d0\u51fa\u7684\u65b9\u6cd5\u7684\u6709\u6548\u6027\u3002<\/span><\/section>\n


<\/h2>\n

<\/section>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u4e09\uff08Theis Lange\uff0cChair\uff09<\/strong><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Continuous-time TMLE for multivariate causal parameters in time-to-event settings<\/span><\/strong><\/span><\/h4>\n
Helene Charlotte Wiese Rytgaard<\/span><\/section>\n
\u76ee\u6807\u5b66\u4e60\uff08TMLE\uff09\u662f\u4e00\u79cd\u5c06\u673a\u5668\u5b66\u4e60\u4e0e\u6e10\u8fd1\u7edf\u8ba1\u63a8\u65ad\u76f8\u7ed3\u5408\u7684\u56e0\u679c\u53c2\u6570\u534a\u53c2\u6709\u6548\u66ff\u4ee3\u4f30\u8ba1\u7684\u4e00\u822c\u65b9\u6cd5\u3002\u8fde\u7eed\u65f6\u95f4TMLE\u662f\u5728\u7eb5\u5411\u6570\u636e\u4e2d\u76ee\u6807\u5b66\u4e60\u65b9\u6cd5\u7684\u4e00\u822c\u5316\uff0c\u5728\u7eb5\u5411\u6570\u636e\u8bbe\u5b9a\u4e2d\uff0c\u5e72\u9884\u3001\u534f\u53d8\u91cf\u548c\u7ed3\u679c\u53ef\u4ee5\u53d1\u751f\u5728\u4efb\u4f55\u7279\u5b9a\u7684\u65f6\u95f4\u70b9\u3002\u62a5\u544a\u4eba\u8003\u8651\u4e86\u7ecf\u5178\u7684\u65f6\u95f4-\u4e8b\u4ef6\u8bbe\u5b9a\u2014\u2014\u6cbb\u7597\u524d\u7684\u534f\u53d8\u91cf\u53ef\u80fd\u5bf9\u57fa\u7ebf\u6cbb\u7597\u51b3\u7b56\u548c\u7ed3\u5c40\u8d77\u6df7\u6742\u4f5c\u7528\uff0c\u5e76\u7279\u522b\u5173\u6ce8\u5728\u4e0d\u540c\u65f6\u95f4\u548c\u4e8b\u4ef6\u7c7b\u578b\u4e0b\u5982\u4f55\u540c\u65f6\u4f30\u8ba1\u751f\u5b58\u6982\u7387\u548c\u7edd\u5bf9\u98ce\u9669\u6982\u7387\u3002\u6240\u63d0\u51fa\u7684\u4f30\u8ba1\u91cf\u4e0d\u4ec5\u662f\u6e10\u8fd1\u7ebf\u6027\u3001\u6709\u6548\u3001\u9075\u5faa\u7531\u975e\u53c2\u6570\u6709\u6548\u5f71\u54cd\u51fd\u6570\u5b8c\u5168\u8868\u793a\u7684\u6e10\u8fd1\u5206\u5e03\uff0c\u800c\u4e14\u8fd8\u4fdd\u8bc1\u4e86\u53c2\u6570\u7a7a\u95f4\u7684\u7ea6\u675f\u6761\u4ef6\uff0c\u5982\u5355\u8c03\u6027\u4ee5\u53ca\u6982\u7387\u548c\u4e3a1\u3002<\/span><\/section>\n

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

Drop-in of concomitant medication<\/span><\/strong><\/span><\/h4>\n
Edwin Fong<\/span><\/section>\n
\u4e34\u5e8a\u8bd5\u9a8c\u4e2d\u7684\u968f\u673a\u5316\u786e\u4fdd\u4e86\u6cbb\u7597\u5206\u914d\u65e0\u6df7\u6742\uff0c\u4f46\u968f\u673a\u5316\u540e\u53d1\u751f\u7684\u5e76\u53d1\u4e8b\u4ef6\u53ef\u80fd\u662f\u4e0d\u5e73\u8861\u7684\u3002\u5728\u7cd6\u5c3f\u75c5\u4e34\u5e8a\u8bd5\u9a8c\u4e2d\uff0c\u4f34\u968f\u7528\u836f\u662f\u4e00\u4e2a\u5173\u952e\u7684\u5e76\u53d1\u4e8b\u4ef6\u3002\u62a5\u544a\u4eba\u8ba4\u4e3a\u4f20\u7edf\u7684\u611f\u5174\u8da3\u7684\u91cf\u548c\u4f30\u8ba1\u65b9\u6cd5\uff0c\u4e0d\u80fd\u76f4\u63a5\u5904\u7406\u6709\u4f34\u968f\u7528\u836f\u65f6\u7684\u4e34\u5e8a\u8bd5\u9a8c\u3002\u56e0\u6b64\u63d0\u51fa\u4e86\u66f4\u9002\u5408\u8be5\u60c5\u5f62\u7684\u88ab\u4f30\u91cf\uff0c\u5e76\u5728\u5e8f\u8d2f\u53ef\u5ffd\u7565\u6027\u5047\u8bbe\u4e0b\u4f7f\u7528\u7eb5\u5411\u76ee\u6807\u6700\u5927\u4f3c\u7136\u4f30\u8ba1(LTMLE)\u6765\u5904\u7406\u6cbb\u7597\u7ec4\u548c\u5b89\u6170\u5242\u7ec4\u4e4b\u95f4\u4f34\u968f\u7528\u836f\u7684\u4e0d\u5e73\u8861\u3002\u8be5\u65b9\u6cd5\u53ef\u4ee5\u62d3\u5c55\u5230\u5b58\u5728\u7ade\u4e89\u98ce\u9669\u548c\u53f3\u5220\u5931\u7684\u60c5\u5f62\uff0c\u5e76\u5728\u6a21\u62df\u6570\u636e\u4e2d\u53d6\u5f97\u4e86\u8f83\u597d\u6548\u679c\u3002<\/span><\/section>\n

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

A Cross-Validated Targeted Maximum Likelihood Estimator for Data-Adaptive Experiment Selection Applied to the Augmentation of RCT Control Arms with Observational Data<\/span><\/strong><\/span><\/h4>\n
Lauren Dang<\/span><\/section>\n
\u5f53\u65e0\u6cd5\u8fdb\u884c\u5927\u89c4\u6a21\u7684\u968f\u673a\u4e34\u5e8a\u8bd5\u9a8c\u65f6\uff0c\u4f7f\u7528\u5916\u90e8\u6570\u636e\u589e\u53ef\u80fd\u4f1a\u589e\u52a0\u529f\u6548\uff0c\u4f46\u6709\u4ea7\u751f\u504f\u501a\u7684\u98ce\u9669\u3002\u73b0\u6709\u7684\u7ec4\u5408\u6570\u636e\u6765\u6e90\u7684\u65b9\u6cd5\u901a\u5e38\u4f9d\u8d56\u4e8e\u4e25\u683c\u7684\u5047\u8bbe\uff0c\u6216\u8005\u5728\u5b58\u5728\u504f\u501a\u7684\u60c5\u51b5\u4e0b\u53ef\u80fd\u53cd\u800c\u964d\u4f4e\u529f\u6548\u3002\u62a5\u544a\u4eba\u63d0\u51fa\u4e86\u4ea4\u53c9\u9a8c\u8bc1\u7684\u76ee\u6807\u6700\u5927\u4f3c\u7136\u4f30\u8ba1(CV-TMLE)\u65b9\u6cd5\u6765\u6570\u636e\u81ea\u9002\u5e94\u5730\u9009\u62e9\u6700\u4f18\u65b9\u6848\u2014\u2014\u53ea\u7528RCT\uff08\u5982\u679c\u4e0d\u5b58\u5728\u65e0\u504f\u5916\u90e8\u6570\u636e\uff09\u6216\u7528RCT\u548c\u5916\u90e8\u6570\u636e\u3002\u7b97\u6cd5\u5c06\u7ecf\u9a8c\u5206\u5e03\u6620\u5c04\u5230\u65b9\u6848\u9009\u62e9\u5668\u4e2d\uff0c\u8be5\u9009\u62e9\u5668\u53ef\u4f18\u5316\u611f\u5174\u8da3\u7684\u56e0\u679c\u6548\u5e94\u7684\u504f\u5dee-\u65b9\u5dee\u6743\u8861\u3002\u62a5\u544a\u4eba\u4f7f\u7528\u6a21\u62dfRCT\u6570\u636e\u548c\u6709\u4e0d\u540c\u7a0b\u5ea6\u504f\u5dee\u7684\u5916\u90e8\u6570\u636e\uff0c\u4f30\u8ba1\u5e73\u5747\u6cbb\u7597\u6548\u679c\u3002CV-TMLE\u76f8\u6bd4\u5176\u4ed6\u65b9\u6cd5\uff08\u5305\u62ec\u4ec5RCT\uff09\u76f8\u6bd4\u8986\u76d6\u7387\u63a5\u8fd1\uff0c\u800c\u529f\u6548\u660e\u663e\u589e\u52a0\u3002<\/span><\/section>\n

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

Generalisability and transportability in the context of target trial emulations<\/span><\/strong><\/span><\/h4>\n
Kim Katrine Bjerring Clemmense<\/span><\/section>\n
2016\u5e74\uff0cHernan\u548cRobins\u6b63\u5f0f\u63d0\u51fa\u4e86\u89c2\u6d4b\u6570\u636e\u5206\u6790\u7684\u76ee\u6807\u8bd5\u9a8c\u65b9\u6cd5\u3002\u76ee\u6807\u8bd5\u9a8c\u662f\u6211\u4eec\u4e3a\u4e86\u56de\u7b54\u6211\u4eec\u611f\u5174\u8da3\u7684\u95ee\u9898\u800c\u8fdb\u884c\u7684\u5047\u60f3\u968f\u673a\u8bd5\u9a8c\u3002\u4f7f\u7528\u8fd9\u79cd\u65b9\u6cd5\uff0c\u89c2\u5bdf\u6570\u636e\u7684\u5206\u6790\u53ef\u4ee5\u88ab\u89c6\u4e3a\u6a21\u62df\u76ee\u6807\u8bd5\u9a8c\u7684\u4e00\u79cd\u5c1d\u8bd5\u3002\u62a5\u544a\u4eba\u7b80\u8981\u4ecb\u7ecd\u4e86\u76ee\u6807\u5b9e\u9a8c\u6cd5\uff0c\u5e76\u4e3e\u4f8b\u8bf4\u660e\u4e86\u8be5\u65b9\u6cd5\u5982\u4f55\u7528\u4e8e\u968f\u673a\u5bf9\u7167\u8bd5\u9a8c\u7ed3\u679c\u7684\u53ef\u6cdb\u5316\u6027\u548c\u53ef\u79fb\u690d\u6027\u95ee\u9898\u3002<\/span><\/section>\n


<\/h2>\n


<\/p>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u56db\uff08Lu Wang\uff0cChair\uff09<\/strong><\/span><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Controlling weak instrument bias in multivariable Mendelian randomization using empirical shrinkage<\/span><\/strong><\/span><\/h4>\n
Jean Morrison<\/span><\/section>\n
\u6559\u6388\u9996\u5148\u56de\u987e\u4e86\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u7814\u7a76\u4e2d\u5229\u7528\u6c47\u603b\u6570\u636e\u7684\u4f20\u7edf\u65b9\u6cd5\uff0c\u4f8b\u5982\u9006\u65b9\u5dee\u52a0\u6743\u65b9\u6cd5\u548c\u6781\u5927\u4f3c\u7136\u65b9\u6cd5\u7b49\u3002\u5f53\u5b58\u5728\u57fa\u56e0\u591a\u6548\u6027\u65f6\uff0c\u5df2\u6709\u7684\u65b9\u6cd5\u6709Egger\u56de\u5f52\uff0c\u4e2d\u4f4d\u6570\u56de\u5f52\u7b49\u3002\u7136\u540e\u6559\u6388\u63d0\u51fa\u505a\u591a\u5143\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u7814\u7a76\u3002\u8be5\u65b9\u6cd5\u53ef\u4ee5\u540c\u65f6\u7814\u7a76\u591a\u4e2a\u6027\u72b6\u5bf9\u4e8e\u75be\u75c5\u7684\u5f71\u54cd\u3002\u7136\u800c\u5f53\u6027\u72b6\u8fc7\u591a\u65f6\uff0c\u5728\u9009\u62e9\u9700\u8981\u63a7\u5236\u7684\u6df7\u6742\u53d8\u91cf\u65f6\uff0c\u4f1a\u5bfc\u81f4\u6700\u540e\u7684\u4f30\u8ba1\u4ea7\u751f\u504f\u5dee\u3002\u4e3a\u4e86\u89e3\u51b3\u8be5\u95ee\u9898\uff0c\u6559\u6388\u63d0\u51fa\u4e86\u57fa\u4e8e\u53d8\u5206\u8d1d\u53f6\u65af\u7b97\u6cd5\u7684\u8ba1\u7b97\u65b9\u6cd5\uff08ESMR\uff09\uff0c\u4e0e\u4e4b\u524d\u7684\u9006\u6982\u7387\u52a0\u6743\u65b9\u6cd5\u548cGRAPPLE\u65b9\u6cd5\u76f8\u6bd4\u8f83\uff0c\u65b0\u65b9\u6cd5\u80fd\u591f\u63a7\u5236\u504f\u5dee\u5728\u5c0f\u8303\u56f4\u5185\uff0c\u540c\u65f6\u8ba1\u7b97\u65f6\u95f4\u8f83\u77ed\uff0c\u8be5\u65b0\u65b9\u6cd5\u5728\u6a21\u62df\u4e2d\u5c55\u793a\u51fa\u8f83\u597d\u7684\u6027\u8d28\u3002<\/span><\/section>\n

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

A Bayesian Reinforcement Learning Framework for Optimizing Sequential Combination Antiretroviral Therapy in People with HIV<\/span><\/strong><\/span><\/h4>\n
Yanxun Xu<\/span><\/section>\n
\u7531\u4e8e\u5728HIV\u6cbb\u7597\u4e2d\u7684\u9700\u8981\uff0c\u5e8f\u8d2f\u51b3\u7b56\u95ee\u9898\u5177\u6709\u5f88\u5f3a\u7684\u5b9e\u7528\u4ef7\u503c\u3002\u4f46\u662f\u5728\u4f30\u8ba1\u51b3\u7b56\u7684\u56e0\u679c\u4f5c\u7528\u65f6\uff0c\u6709\u8bb8\u591a\u6280\u672f\u4e0a\u7684\u56f0\u96be\u3002\u9996\u5148\u9ad8\u7ef4\u7684\u72b6\u6001\u7a7a\u95f4\u5f88\u96be\u4f30\u8ba1\uff0c\u6570\u636e\u4e0d\u5e73\u8861\u95ee\u9898\u4e5f\u5341\u5206\u4e25\u91cd\uff0c\u6b64\u5916\u5982\u4f55\u4ece\u5de8\u5927\u7684\u51b3\u7b56\u7a7a\u95f4\u4e2d\u5f97\u5230\u4e00\u4e2a\u5b9e\u7528\u7684\u51b3\u7b56\u65b9\u6848\u4e5f\u96be\u4ee5\u89e3\u51b3\u3002\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e9b\u95ee\u9898\uff0c\u8bb8\u6559\u6388\u63d0\u51fa\u5229\u7528\u8d1d\u53f6\u65af\u52a8\u6001\u6a21\u578b\u6765\u62df\u5408\u6574\u4e2a\u52a8\u6001\u8fc7\u7a0b\uff0c\u7136\u540e\u5229\u7528\u60e9\u7f5a\u9879\u6765\u5904\u7406\u73af\u5883\u4e2d\u7684\u9650\u5236\u6761\u4ef6\uff0c\u5728\u4f7f\u7528\u7ebf\u4e0b\u5f3a\u5316\u5b66\u4e60\u65b9\u6cd5\u4f18\u5316\u76ee\u6807\u51fd\u6570\uff0c\u6700\u7ec8\u5f97\u5230\u9700\u8981\u7684\u5e8f\u8d2f\u51b3\u7b56\u65b9\u6848\u3002\u8be5\u65b9\u6cd5\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u5bf9\u4e8e\u7597\u6548\u6709\u4e86\u663e\u8457\u7684\u63d0\u5347\u3002<\/span><\/section>\n

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

Constructing stabilized dynamic surveillance rules for optimal monitoring schedules<\/span><\/strong><\/span><\/h4>\n
Yingqi Zhao<\/span><\/section>\n
\u524d\u5217\u817a\u764c\u662f\u7537\u6027\u4e2d\u7b2c\u4e8c\u5e38\u89c1\u7684\u764c\u75c7\uff0c\u4e3b\u52a8\u76d1\u6d4b\u5728\u5176\u4e2d\u53d1\u6325\u91cd\u8981\u4f5c\u7528\u3002\u8d75\u6559\u6388\u5b9a\u4e49\u4e86\u4e00\u79cd\u52a8\u6001\u76d1\u6d4b\u89c4\u5219\uff08Dynamic Surveillance Rule, DSR\uff09\uff0c\u5e76\u7ed3\u5408\u771f\u9633\u6027\u5206\u6570\u4e0e\u5047\u9633\u6027\u5206\u6570\u4f5c\u4e3a\u8bc4\u4ef7\u6807\u51c6\u3002\u8d75\u6559\u6388\u4e3b\u8981\u5173\u6ce8\u7a33\u5b9a\u7ebf\u6027\u89c4\u5219\uff0c\u8be5\u7c7b\u89c4\u5219\u6709\u4e09\u4e2a\u7279\u70b9\uff1a\u4e00\u662f\u89c4\u5219\u516c\u5f0f\u4e0d\u968f\u65f6\u95f4\u53d8\u5316\uff1b\u4e8c\u662f\u4e2a\u4f53\u662f\u5426\u88ab\u5efa\u8bae\u76d1\u6d4b\u53ef\u56e0\u968f\u65f6\u95f4\u53d8\u5316\u7684\u534f\u53d8\u91cf\u53d8\u5316\u800c\u53d8\u5316\uff1b\u4e09\u662f\u5728\u5b9e\u9645\u4e2d\u5bb9\u6613\u5b9e\u73b0\u3002\u5728\u6b64\u57fa\u7840\u4e0a\uff0c\u8d75\u6559\u6388\u7ed9\u51fa\u4e24\u79cd\u7b97\u6cd5\uff0c\u4e00\u662f\u5171\u4eab\u5efa\u6a21\uff08Shared modeling\uff09\u6cd5\uff0c\u4e8c\u662f\u4f18\u5316\u4ee3\u7406\u51fd\u6570\uff08Optimization with surrogate function\uff09\u6cd5\u3002\u6700\u540e\u6a21\u62df\u6570\u636e\u548cCanary PASS\u6570\u636e\u9a8c\u8bc1\u4e86\u65b9\u6cd5\u7684\u6709\u6548\u6027\u3002<\/span><\/section>\n

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


<\/h2>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u4e94\uff08Robin Evans\uff0cChair\uff09<\/strong><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

On structural imsets for describing and learning graphical models<\/span><\/strong><\/span><\/h4>\n
Milan Studen\u00fd   Academy of Sciences of the Czech Republic<\/span><\/section>\n
Studen\u00fd\u6559\u6388\u7684\u6f14\u8bb2\u662f\u7ed3\u6784\u6027imsets\u65b9\u6cd5\u7684\u7b80\u8981\u56de\u987e\uff0c\u8be5\u65b9\u6cd5\u53ef\u4ee5\u7528\u6765\u63cf\u8ff0\u79bb\u6563\u968f\u673a\u53d8\u91cf\u7684\u6761\u4ef6\u72ec\u7acb\u7ed3\u6784\u3002Imsets\u662f\u7531N\u4e2a\u53d8\u91cf\u7684\u57fa\u672c\u96c6\u5408\u7684\u5b50\u96c6\u7f16\u7801\u7684\u6574\u6570\u7ec4\u6210\u7684\u7279\u6b8a\u5411\u91cf\u3002\u8bb2\u5ea7\u91cd\u70b9\u5f3a\u8c03\u4e86\u63cf\u8ff0\u6761\u4ef6\u72ec\u7acb\u7ed3\u6784\u7684\u56fe\u6a21\u578b\uff0c\u5f15\u51fa\u5e76\u4ecb\u7ecd\u4e86\u6574\u6570\u7ebf\u6027\u89c4\u5212\u7684\u65b9\u6cd5\u6765\u5b66\u4e60\u53ef\u5206\u89e3\u7684\u56fe\u6a21\u578b\u7684\u7ed3\u6784\u3002<\/span><\/section>\n

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

Towards standard imsets for maximal ancestral graphs<\/span><\/strong><\/span><\/h4>\n
Zhongyi Hu    Department of Statistics, University of Oxford<\/span><\/section>\n
Imset\u662f\u8868\u793a\u6761\u4ef6\u72ec\u7acb\u6a21\u578b\u7684\u4ee3\u6570\u65b9\u6cd5\u3002\u5728\u6761\u4ef6\u72ec\u7acb\u6a21\u578b\u4e2d\uff0cImsets\u6709\u5438\u5f15\u4eba\u7684\u6027\u8d28\uff0c\u5e76\u4e14\u5f88\u9002\u5408\u4e0e\u6709\u5411\u65e0\u73af\u56fe\u4e00\u8d77\u4f7f\u7528\u3002\u7279\u522b\u5730\uff0c\u4e00\u4e2a\u6709\u5411\u65e0\u73af\u56fe\u7684\u6807\u51c6\u7684imset\u548c\u5b83\u6240\u8868\u793a\u7684\u72ec\u7acb\u6027\u7ed3\u6784\u662f\u4e00\u4e00\u5bf9\u5e94\u7684\uff0c\u56e0\u6b64\u6807\u51c6\u7684imset\u4e5f\u662f\u9a6c\u5c14\u79d1\u592b\u7b49\u4ef7\u7c7b\u7684\u6807\u7b7e\u3002<\/span><\/section>\n
\u80e1\u4ef2\u6613\u8001\u5e08\u9488\u5bf9\u6700\u5927\u7956\u5148\u56fe\uff08MAG\uff09\uff0c\u4f7f\u7528\u201cparameterizing set representation\u201d\u62d3\u5c55\u4e86\u6807\u51c6\u7684imset\uff0c\u8fd9\u6837\uff0c\u6784\u9020\u51fa\u7684imset\u4e5f\u80fd\u591f\u4ee3\u8868\u6700\u5927\u7956\u5148\u56fe\u7684\u9a6c\u5c14\u79d1\u592b\u7b49\u4ef7\u7c7b\u3002\u5bf9\u5f88\u591a\u56fe\u6765\u8bf4\uff0c\u65b0\u63d0\u51fa\u7684imset\u662f\u6ee1\u8db3\u9a6c\u5c14\u53ef\u592b\u6027\u7684\uff0c\u56e0\u6b64\u80fd\u591f\u901a\u8fc7\u5ea6\u91cf\u4e00\u7cfb\u5217\u7684\u72ec\u7acb\u6027\u5dee\u5f02\u6765\u63d0\u4f9b\u4e00\u79cd\u8bc4\u5206\u51c6\u5219\uff0c\u8fd9\u5c31\u7ed9\u4e86\u6211\u4eec\u4e00\u79cdBIC\u8bc4\u5206\u4ee5\u5916\u7684\u9009\u62e9\u3002\u7136\u800c\uff0c\u5728\u6709\u4e9b\u6a21\u578b\u4e0b\u8fd9\u4e00\u8868\u5f81\u5e76\u4e0d\u6210\u7acb\uff0c\u5e76\u4e14\u5728\u4e00\u4e9b\u60c5\u51b5\u4e0b\u4e5f\u4e0d\u80fd\u4ee3\u8868\u4efb\u4f55\u72ec\u7acb\u6027\uff0c\u56e0\u6b64\uff0c\u80e1\u4ef2\u6613\u8001\u5e08\u63a5\u7740\u8bc1\u660e\u4e86\u5728\u4e00\u4e9b\u7b80\u5355\u7684\u56fe\u4e0b\u8fd9\u4e00\u8868\u5f81\u786e\u5b9e\u6210\u7acb\uff0c\u5e76\u4e14\u662f\u6700\u7b80\u5355\u7684\u4e00\u79cd\u3002\u6700\u540e\uff0c\u6f14\u8bb2\u4ecb\u7ecd\u4e86\u6709\u5e8f\u5c40\u90e8\u9a6c\u5c14\u79d1\u592b\u6027\u7684\u6539\u8fdb\u65b9\u6cd5\uff0c\u5e76\u4e14\u7528\u5b83\u6765\u627e\u5230\u80fd\u591f\u8868\u793a\u6700\u5927\u7956\u5148\u56fe\u7684\u6700\u4f73\u7684imset\u3002<\/span><\/section>\n


<\/h4>\n

Using imsets to score causal models with latent confounding<\/span><\/strong><\/span><\/h4>\n
Bryan Andrews  Department of Philosophy, Carnegie Mellon University<\/span><\/section>\n
\u5728\u7edf\u8ba1\u548c\u673a\u5668\u5b66\u4e60\u4e2d\uff0c\u6709\u5411\u65e0\u73af\u56fe\uff08DAG\uff09\u88ab\u5e7f\u6cdb\u5730\u7814\u7a76\u548c\u5e94\u7528\uff0c\u5b83\u4eec\u662f\u5b66\u4e60\u548c\u63a8\u65ad\u7684\u7b80\u4fbf\u6709\u6548\u7684\u65b9\u6cd5\u3002\u7136\u800c\uff0c\u8fd9\u4e00\u7c7b\u6a21\u578b\u5e76\u4e0d\u80fd\u5f88\u597d\u5730\u5904\u7406\u6709\u6f5c\u5728\u7684\u6df7\u6742\u7684\u7cfb\u7edf\u3002\u6709\u5411\u65e0\u73af\u6df7\u5408\u56fe\u6a21\u578b\uff08ADMG\uff09\u63cf\u8ff0\u4e86\u6709\u5411\u65e0\u73af\u56fe\u6a21\u578b\u7684\u8fb9\u7f18\u7279\u5f81\uff0c\u56e0\u6b64\u4ed6\u4eec\u66f4\u9002\u7528\u4e8e\u5904\u7406\u8fd9\u4e00\u7c7b\u6709\u6f5c\u5728\u6df7\u6742\u7684\u7cfb\u7edf\u3002\u7136\u800c\uff0cADMG\u6a21\u578b\u56e0\u5176\u590d\u6742\u6027\u800c\u81f3\u4eca\u6ca1\u6709\u5f97\u5230\u5e7f\u6cdb\u7684\u5e94\u7528\u3002<\/span><\/section>\n
\u5728Andrews\u6559\u6388\u7684\u8bb2\u5ea7\u4e2d\uff0c\u4ed6\u63a2\u8ba8\u4e86\u5c06\u6709\u5411\u65e0\u73af\u56fe\u6a21\u578b\u7684\u7279\u5f81imset\u62d3\u5c55\u5230\u6709\u5411\u65e0\u73af\u6df7\u5408\u56fe\u6a21\u578b\u7684\u65b9\u6cd5\u3002\u8fdb\u4e00\u6b65\uff0cAndrews\u6559\u6388\u8ba8\u8bba\u4e86\u8fd9\u4e00\u62d3\u5c55\u540e\u7684ADMG\u6a21\u578b\u7684\u5206\u89e3\u6807\u51c6\uff0c\u4ee5\u53ca\u5b83\u4e0e\u5168\u5c40\u9a6c\u5c14\u79d1\u592b\u6027\u8d28\u7684\u7b49\u4ef7\u6027\u3002\u6700\u540e\uff0c\u8bb2\u5ea7\u5c55\u793a\u4e86\u4e00\u4e2a\u76f8\u5408\u5f97\u5206\u51c6\u5219\uff0c\u5e76\u5c06\u5176\u7528\u4e8e\u5b66\u4e60ADMG\uff0c\u4ee5\u5c55\u793a\u5206\u89e3\u65b9\u6cd5\u7684\u5b9e\u7528\u6027\u3002<\/span><\/section>\n


<\/h4>\n

Imsets and supermodular functions<\/span><\/strong><\/span><\/h4>\n
James Cussens   Department of Computer Science, University of Bristol<\/span><\/section>\n
Supermodular\u51fd\u6570\u4e3a\u63cf\u8ff0\u6761\u4ef6\u72ec\u7acb\u7ed3\u6784\uff0c\u5728imsets\u4ee5\u5916\uff0c\u63d0\u4f9b\u4e86\u4e00\u79cd\u4e8c\u5143\u7684\u66ff\u4ee3\u65b9\u6cd5\u3002\u5728\u8fd9\u4e00\u8bb2\u5ea7\u4e2d\uff0cCussens\u6559\u6388\u8bb2\u8ff0\u4e86imsets\u548csupermodular\u51fd\u6570\u4e4b\u95f4\u7684\u5173\u7cfb\uff0c\u63a2\u7d22\u4e86\u7528supermodular\u51fd\u6570\u6765\u63cf\u8ff0\u6761\u4ef6\u72ec\u7acb\u5173\u7cfb\u7684\u5229\u5f0a\u3002\u4e00\u65b9\u9762\uff0c\u5bf9\u4e00\u4e2a\u6761\u4ef6\u72ec\u7acb\u5173\u7cfb\u6765\u8bf4\uff0c\u76f8\u6bd4\u4e8eimset\uff0c\u7528supermodular\u51fd\u6570\u6765\u68c0\u9a8c\u8fd9\u4e00\u5173\u7cfb\u662f\u5426\u6210\u7acb\u662f\u66f4\u52a0\u5bb9\u6613\u7684\u3002\u6b64\u5916\uff0c\u8ba1\u7b97\u8fb9\u7f18supermodular\u51fd\u6570\u4e5f\u66f4\u52a0\u7b80\u5355\u3002\u53e6\u4e00\u65b9\u6cd5\uff0c\u73b0\u5982\u4eca\u8fd8\u6ca1\u6709\u4e00\u79cd\u6807\u51c6\u7684supermodular\u51fd\u6570\u80fd\u591f\u7528\u6765\u4ee3\u8868\u8d1d\u53f6\u65af\u7f51\u7edc\u7ed3\u6784\uff0c\u7136\u800cimset\u6709\u3002\u4efb\u4f55\u4e00\u4e2a\u62df\u9635\u7684\u79e9\u51fd\u6570\u5373\u662f\u4e00\u4e2asubmodular\u51fd\u6570\uff0c\u56e0\u6b64\u62df\u9635\u53ef\u4ee5\u4e3a\u7279\u5b9a\u7684supermodular\u51fd\u6570\u63d0\u4f9b\u7d27\u7684\u8868\u793a\u3002Cussens\u6559\u6388\u7814\u7a76\u4e86\u5982\u4f55\u63a2\u7d22\u8fd9\u4e00supermodular\u51fd\u6570\u4e0e\u62df\u9635\u7684\u5173\u8054\u3002<\/span><\/section>\n

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


<\/h2>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u516d\uff08Xiao-Hua Zhou\uff0cChair\uff09<\/strong><\/span><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Total causal effects in MPDAGs: identification and minimal enumeration<\/span><\/strong><\/span><\/h4>\n
Emilija Perkovic, University of Washington<\/span><\/section>\n
\u5f53\u56e0\u679c\u56fe\u4e0d\u5b8c\u5168\u5df2\u77e5\u65f6\uff0c\u603b\u56e0\u679c\u4f5c\u7528\u53ef\u80fd\u65e0\u6cd5\u8bc6\u522b\u3002\u62a5\u544a\u4eba\u8003\u8651\u6700\u5927\u5b9a\u5411\u6709\u5411\u65e0\u73af\u56fe\uff08Maximally oriented partially directed acyclic graph, MPDAG\uff09\uff0c\u5176\u4e2d\u6709\u4e00\u4e9b\u8fb9\u7684\u65b9\u5411\u672a\u77e5\u3002\u76ee\u524d\u5728\u56e0\u679c\u56fe\u4e0a\u8bc6\u522b\u56e0\u679c\u4f5c\u7528\u7684\u65b9\u5f0f\u6709\u4e24\u79cd\uff1a\u4e00\u662f\u8c03\u6574\u89c4\u5219\uff0c\u5b83\u662f\u5728DAG\u6216MPDAG\u4e0a\u8bc6\u522b\u56e0\u679c\u4f5c\u7528\u7684\u5145\u5206\u6761\u4ef6\uff1b\u4e8c\u662fG-\u516c\u5f0f\u5373\u622a\u65ad\u5206\u89e3\uff0c\u5b83\u662f\u5728DAG\u4e0a\u8bc6\u522b\u56e0\u679c\u4f5c\u7528\u7684\u5145\u5206\u5fc5\u8981\u6761\u4ef6\uff0c\u4f46\u5c1a\u672a\u63a8\u5e7f\u5230MPDAG\u4e0a\u3002\u62a5\u544a\u4eba\u63d0\u51fa\u4e86\u5728DAG\u6216MPDAG\u4e0a\u8bc6\u522b\u56e0\u679c\u4f5c\u7528\u7684\u5145\u5206\u5fc5\u8981\u6761\u4ef6\uff1a\u5904\u7406\u53d8\u91cf\u5230\u7ed3\u5c40\u53d8\u91cf\u7684\u6b63\u5e38\u56e0\u679c\u8def\u5f84\u90fd\u4ee5\u76f4\u63a5\u8fb9\u5f00\u59cb\u3002\u5728\u6b64\u6761\u4ef6\u4e0b\uff0c\u62a5\u544a\u4eba\u7ed9\u51fa\u4e86\u56e0\u679c\u4f5c\u7528\u7684\u622a\u65ad\u5206\u89e3\u516c\u5f0f\u3002\u53e6\u5916\uff0c\u5f53\u56e0\u679c\u4f5c\u7528\u4e0d\u53ef\u8bc6\u522b\u65f6\uff0c\u62a5\u544a\u4eba\u7ed9\u51fa\u4e86\u901a\u8fc7\u9012\u5f52\u679a\u4e3e\u6240\u6709\u53ef\u80fd\u56e0\u679c\u4f5c\u7528\u7684\u7b97\u6cd5\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Perkovic, E. (2020). Identifying causal effects in maximally oriented partially directed acyclic graphs. In Conference on Uncertainty in Artificial Intelligence (pp. 530-539). PMLR.<\/span><\/section>\n


<\/h4>\n

A simple asymptotic non-conservative test for indirect effect<\/span><\/strong><\/span><\/h4>\n
Gary Chan, University of Washington<\/span><\/section>\n
\u5728\u4e2d\u4ecb\u5206\u6790\u4e2d\uff0c\u5904\u7406\u53d8\u91cf\u53ef\u901a\u8fc7\u76f4\u63a5\u8def\u5f84\u548c\u95f4\u63a5\u8def\u5f84\u5f71\u54cd\u7ed3\u5c40\u3002\u6709\u65f6\u6211\u4eec\u5e0c\u671b\u68c0\u9a8c\u95f4\u63a5\u4f5c\u7528\u662f\u5426\u5b58\u5728\u3002\u5728\u7ebf\u6027\u7ed3\u6784\u65b9\u7a0b\u6a21\u578b\u4e2d\uff0c\u5bf9\u95f4\u63a5\u4f5c\u7528\u7684\u68c0\u9a8c\u53ef\u4ee5\u5199\u6210H(0):\u03b8(1)\u03b8(2)=0\u3002\u8fd9\u662f\u4e00\u4e2a\u590d\u5408\u68c0\u9a8c\uff0c\u56e0\u4e3a\u5728\u539f\u5047\u8bbe\u4e0b\u03b8(1)\u6216\u03b8(2)\u90fd\u53ef\u80fd\u7b49\u4e8e0\u3002Sobel\u68c0\u9a8c\u662fWald\u7c7b\u578b\u7684\u68c0\u9a8c\uff0c\u5f53\u03b8(1)\u548c\u03b8(2)\u4e8c\u8005\u90fd\u4e3a0\u65f6\u68c0\u9a8c\u7edf\u8ba1\u91cf\u7684\u6e10\u8fd1\u65b9\u5dee\u4f1a\u8fc7\u5c0f\u3002\u8bb0\u68c0\u9a8c\u03b8(1)=0\u548c\u03b8(2)=0\u7684p\u503c\u5206\u522b\u4e3ap(1)\u548cp\u00ad(2)\uff0c\u5728\u6b63\u6001\u6a21\u578b\u4e0b\uff0c\u8be5\u68c0\u9a8c\u7684\u4f3c\u7136\u6bd4\u68c0\u9a8c\u8981\u6c42\u5f53p(1)\u548cp(2)\u7684\u6700\u5927\u503c\u8f83\u5c0f\u65f6\u62d2\u7edd\u539f\u5047\u8bbe\u3002\u62a5\u544a\u4eba\u63d0\u51fa\u4e86\u57fa\u4e8ep\u503c\u7684\u68c0\u9a8c\uff0c\u4f7f\u5f97\u5728\u539f\u5047\u8bbe\u4e0b\uff0c\u7edf\u8ba1\u91cf\u6536\u655b[0,1]\u4e0a\u7684\u5747\u5300\u5206\u5e03\uff0c\u539f\u5047\u8bbe\u4e0d\u6210\u7acb\u65f6\u7edf\u8ba1\u91cf\u6536\u655b\u52300\uff0c\u65b0\u7edf\u8ba1\u91cf\u53ef\u88ab\u770b\u4f5c\u662f\u539f\u59cb\u590d\u5408\u68c0\u9a8c\u7684p\u503c\u3002\u65b0\u68c0\u9a8c\u4e0d\u4f1a\u9047\u5230\u7b2c\u4e00\u7c7b\u9519\u8bef\u7387\u8fc7\u5c0f\u7684\u95ee\u9898\uff0c\u5728\u6837\u672c\u91cf\u8f83\u5c0f\u65f6\u6bd4\u73b0\u6709\u65b9\u6cd5\u5177\u6709\u66f4\u9ad8\u7684\u529f\u6548\uff0c\u4f46\u5728\u5c0f\u6837\u672c\u4e0b\u7a0d\u6709\u7b2c\u4e00\u7c7b\u9519\u8bef\u7387\u81a8\u80c0\u3002\u8fdb\u4e00\u6b65\uff0c\u62a5\u544a\u4eba\u628a\u8fd9\u4e00\u601d\u8def\u63a8\u5e7f\u5230\u591a\u4e2a\u53c2\u6570\u4e58\u79ef\u4e3a0\u68c0\u9a8c\u7684\u60c5\u5f62\u3002<\/span><\/section>\n


<\/h4>\n

Testing an Elaborate Theory of a Causal Hypothesis<\/span><\/strong><\/span><\/h4>\n
Dylan Small, University of Pennsylvania<\/span><\/section>\n
\u5f53R.A.Fisher\u88ab\u95ee\u53ca\u5728\u89c2\u5bdf\u6027\u7814\u7a76\u4e2d\u5982\u4f55\u624d\u80fd\u6f84\u6e05\u4ece\u5173\u8054\u5230\u56e0\u679c\u5173\u7cfb\u7684\u6b65\u9aa4\u65f6\uff0c\u4ed6\u56de\u7b54\u8bf4\uff1a\u201c\u4f7f\u4f60\u7684\u7406\u8bba\u66f4\u52a0\u8be6\u5c3d\u201d\u2014\u2014\u5728\u6784\u5efa\u4e00\u4e2a\u56e0\u679c\u5047\u8bbe\u65f6\uff0c\u5c3d\u53ef\u80fd\u591a\u5730\u8bbe\u60f3\u5176\u5047\u8bbe\u7684\u4e0d\u540c\u540e\u679c\uff0c\u5e76\u89c4\u5212\u89c2\u5bdf\u6027\u7814\u7a76\u4ee5\u53d1\u73b0\u8fd9\u4e9b\u540e\u679c\u4e2d\u7684\u6bcf\u4e00\u4e2a\u662f\u5426\u6210\u7acb\u3002William Cochran\u79f0\u201c\u8fd9\u79cd\u591a\u9636\u6bb5\u7684\u653b\u51fb\u2026\u2026\u662f\u89c2\u5bdf\u6027\u7814\u7a76\u4e2d\u6700\u6709\u529b\u7684\u6b66\u5668\u4e4b\u4e00\u201d\u3002\u5bf9\u7cbe\u5fc3\u8bbe\u8ba1\u7684\u7406\u8bba\u7684\u5404\u4e2a\u90e8\u5206\u8fdb\u884c\u7edf\u8ba1\u68c0\u9a8c\uff0c\u6709\u52a9\u4e8e\u6f84\u6e05\u56e0\u679c\u5047\u8bbe\u7684\u786e\u8bc1\u7a0b\u5ea6\u3002\u5728\u5b9e\u8df5\u4e2d\uff0c\u56e0\u679c\u5047\u8bf4\u7684\u786e\u8bc1\u7a0b\u5ea6\u662f\u901a\u8fc7\u53e3\u5934\u63cf\u8ff0\u54ea\u79cd\u68c0\u9a8c\u4e3a\u54ea\u79cd\u9884\u6d4b\u63d0\u4f9b\u8bc1\u636e\u6765\u8bc4\u4f30\u7684\u3002\u8fd9\u79cd\u53e3\u5934\u4e0a\u7684\u65b9\u6cd5\u53ef\u80fd\u4f1a\u9519\u8fc7\u5b9a\u91cf\u7684\u6a21\u5f0f\u3002\u62a5\u544a\u4eba\u5f00\u53d1\u4e86\u4e00\u79cd\u5b9a\u91cf\u7684\u65b9\u6cd5\uff0c\u5bf9\u6709\u8bc1\u636e\u652f\u6301\u7684\u8be6\u7ec6\u7406\u8bba\u7684\u6570\u91cf\u8fdb\u884c\u7edf\u8ba1\u63a8\u65ad\u3002\u66f4\u5177\u4f53\u5730\uff0c\u62a5\u544a\u4eba\u4ee5\u5177\u4f53\u6570\u636e\u4e3a\u80cc\u666f\uff0c\u8003\u8651\u4e86\u82e5\u5e72\u4e2a\u72ec\u7acb\u7684\u68c0\u9a8c\uff0c\u4f7f\u7528\u90e8\u5206\u8fde\u63a5\u68c0\u9a8c\u7ed3\u5408\u8fd9\u4e9b\u68c0\u9a8c\u7684\u8bc1\u636e\u3002\u5f53\u5f85\u6bd4\u8f83\u5904\u7406\u7ec4\u548c\u5bf9\u7167\u7ec4\u95f4\u5b58\u5728\u504f\u501a\u65f6\uff0c\u4e5f\u53ef\u4ee5\u6267\u884c\u654f\u611f\u6027\u5206\u6790\uff0c\u6c47\u62a5\u7279\u5b9a\u7a0b\u5ea6\u504f\u501a\u4e0b\u7684\u6700\u5927\u53ef\u80fdp\u503c\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Karmakar, B., & Small, D. S. (2020). Assessment of the extent of corroboration of an elaborate theory of a causal hypothesis using partial conjunctions of evidence factors. The Annals of Statistics, 48(6), 3283-3311.<\/span><\/section>\n


<\/h2>\n


<\/p>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u4e03\uff08Wang Miao\uff0cChair\uff09<\/strong><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

On Mendelian Randomization Mixed-Scale Treatment Effect Robust Identification and Estimation for Causal Inference<\/span><\/strong><\/span><\/h4>\n
Zhonghua Liu, Columbia University<\/span><\/section>\n
\u5728\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u7814\u7a76\u4e2d\uff0c\u5982\u679c\u4f5c\u4e3a\u5de5\u5177\u53d8\u91cf\u7684\u9057\u4f20\u53d8\u4f53\u4e0e\u611f\u5174\u8da3\u7684\u7ed3\u5c40\u5b58\u5728\u6df7\u6742\uff0c\u800c\u4e0d\u662f\u7531\u6cbb\u7597\u53d8\u91cf\u4ecb\u5bfc\u7684\uff0c\u6216\u8005\u5bf9\u611f\u5174\u8da3\u7684\u7ed3\u5c40\u6c34\u5e73\u6709\u591a\u6001\u6027\u5f71\u54cd\uff0c\u90a3\u4e48\u6807\u51c6\u7684\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u5206\u6790\u5c31\u4f1a\u4ea7\u751f\u6709\u504f\u7684\u7ed3\u679c\u3002\u62a5\u544a\u4eba\u901a\u8fc7\u5229\u7528\u4e00\u4e2a\u53ef\u80fd\u65e0\u6548\u7684\u5de5\u5177\u53d8\u91cf\uff0c\u4e3a\u66b4\u9732\u7684\u56e0\u679c\u6548\u5e94\u63d0\u4f9b\u4e86\u65b0\u7684\u8bc6\u522b\u6761\u4ef6\uff0c\u8fd9\u91cc\u5de5\u5177\u53d8\u91cf\u7684\u72ec\u7acb\u6027\u548c\u6392\u9664\u9650\u5236\u5047\u8bbe\u90fd\u53ef\u80fd\u88ab\u8fdd\u53cd\u3002\u63d0\u51fa\u7684\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u6df7\u5408\u5c3a\u5ea6\u6cbb\u7597\u6548\u679c\u7a33\u5065\u8bc6\u522b\uff08MR MiSTERI\uff09\u65b9\u6cd5\u4f9d\u8d56\u4e8e\uff08i\uff09\u5047\u8bbe\u6cbb\u7597\u6548\u679c\u4e0d\u968f\u52a0\u53ef\u5c3a\u5ea6\u4e0a\u53ef\u80fd\u65e0\u6548\u7684\u5de5\u5177\u53d8\u91cf\u800c\u53d8\u5316\uff1b\uff08ii\uff09\u6df7\u6742\u504f\u501a\u4e0d\u968f\u51e0\u7387\u6bd4\u5c3a\u5ea6\u4e0a\u53ef\u80fd\u65e0\u6548\u7684\u5de5\u5177\u53d8\u91cf\u800c\u53d8\u5316\uff1b\u4ee5\u53ca\uff08iii\uff09\u7ed3\u5c40\u7684\u6b8b\u5dee\u76f8\u5bf9\u4e8e\u53ef\u80fd\u65e0\u6548\u7684\u5de5\u5177\u53d8\u91cf\u662f\u5f02\u65b9\u5dee\u7684\u3002\u5c3d\u7ba1\u5047\u8bbe(i)\u548c(ii)\u5df2\u7ecf\u5206\u522b\u51fa\u73b0\u5728\u5de5\u5177\u53d8\u91cf\u6587\u732e\u4e2d\uff0c\u4f46\u5047\u8bbe(iii)\u8fd8\u6ca1\u6709\u51fa\u73b0\uff1b\u6211\u4eec\u6b63\u5f0f\u63d0\u51fa\uff0c\u5373\u4f7f\u6709\u65e0\u6548\u7684\u5de5\u5177\u53d8\u91cf\uff0c\u5b83\u4eec\u7684\u7ed3\u5408\u4e5f\u80fd\u8bc6\u522b\u56e0\u679c\u6548\u5e94\u3002MR MiSTERI\u88ab\u8bc1\u660e\u5728\u5b58\u5728\u53ef\u52a0\u5c3a\u5ea6\u7684\u591a\u56e0\u5b50\u6548\u5e94\u7684\u666e\u904d\u5f02\u8d28\u6027\u65f6\u7279\u522b\u6709\u4f18\u52bf\u3002\u6211\u4eec\u63d0\u51fa\u4e86\u4e00\u4e2a\u7b80\u5355\u800c\u76f8\u5408\u7684\u4e09\u9636\u6bb5\u4f30\u8ba1\u91cf\uff0c\u53ef\u4ee5\u4f5c\u4e3a\u7cbe\u5fc3\u6784\u5efa\u7684\u9ad8\u6548\u4e00\u6b65\u66f4\u65b0\uff08one-step-update\uff09\u4f30\u8ba1\u91cf\u7684\u521d\u6b65\u4f30\u8ba1\u91cf\u3002\u4e3a\u4e86\u7eb3\u5165\u591a\u4e2a\u53ef\u80fd\u76f8\u5173\u7684\u5f31\u65e0\u6548\u5de5\u5177\u53d8\u91cf\u2014\u2014\u8fd9\u662fMR\u7814\u7a76\u4e2d\u5e38\u89c1\u7684\u6311\u6218\uff0c\u62a5\u544a\u4eba\u53ca\u540c\u4e8b\u5f00\u53d1\u4e86\u4e00\u4e2a\u591a\u91cd\u5f31\u65e0\u6548\u5de5\u5177\u53d8\u91cf\uff08MR MaWII MiSTERI\uff09\u65b9\u6cd5\uff0c\u4ee5\u52a0\u5f3a\u8bc6\u522b\u548c\u63d0\u9ad8\u4f30\u8ba1\u51c6\u786e\u5ea6\u3002\u6a21\u62df\u7814\u7a76\u548cUK Biobank\u5206\u6790\u7ed3\u679c\u90fd\u8bc1\u660e\u4e86\u6240\u63d0\u65b9\u6cd5\u7684\u7a33\u5065\u6027\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e<\/span>\uff1a<\/span><\/section>\n
Liu, Z., Ye, T., Sun, B., Schooling, M., & Tchetgen, E. T. (2022). Mendelian randomization mixed\u2010scale treatment effect robust identification and estimation for causal inference. Biometrics.<\/span><\/section>\n


<\/h4>\n

DNN-Based Causal Inference and New Stable Empirical Higher Order Influence Functions<\/span><\/strong><\/span><\/h4>\n
Lin Liu, Shanghai Jiao Tong University<\/span><\/section>\n
\u6df1\u5ea6\u795e\u7ecf\u7f51\u7edc\uff08DNN\uff09\u5df2\u7ecf\u5728\u5404\u79cd\u7edf\u8ba1\u95ee\u9898\u4e0a\u53d6\u5f97\u4e86\u6210\u529f\uff0c\u4ece\u56fe\u50cf\u5206\u7c7b\u5230\u89e3\u51b3\u975e\u7ebf\u6027\u53cd\u95ee\u9898\u3002\u8fd1\u5e74\u6765\uff0c\u4e5f\u6709\u8d8a\u6765\u8d8a\u591a\u7684\u6587\u732e\u7814\u7a76DNN\u5982\u4f55\u5f88\u597d\u5730\u5b66\u4e60\u56e0\u679c\u6548\u5e94\u3002\u5728\u8fd9\u6b21\u6f14\u8bb2\u4e2d\uff0c\u62a5\u544a\u4eba\u8868\u660e\uff0c\u76ee\u524d\u7684\u7406\u8bba\u7ed3\u679c\u5728\u5f88\u5927\u7a0b\u5ea6\u4e0a\u4e0d\u4ee4\u4eba\u6ee1\u610f\uff0c\u53ef\u80fd\u662f\u7531\u4e8eDNN\u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u7684\u9690\u6027\u504f\u501a\u3002\u4f5c\u4e3a\u90e8\u5206\u8865\u6551\u63aa\u65bd\uff0c\u5728Liu\u7b49\u4eba\uff082020\uff09\u7684\u6587\u7ae0\u4e2d\uff0c\u5f00\u53d1\u4e86\u4e00\u4e2a\u57fa\u4e8e\u9ad8\u9636\u5f71\u54cd\u51fd\u6570\uff08HOIF\uff09\u7684\u5047\u8bbe\u68c0\u9a8c\u7a0b\u5e8f\uff0c\u53ef\u4ee5\u63a2\u6d4b\u56e0\u679c\u6548\u5e94\u4f30\u8ba1\u503c\u7684\u504f\u5dee\uff0c\u800c\u4e0d\u53d7\u771f\u5b9e\u6216\u4f30\u8ba1\u7684\u5197\u4f59\u51fd\u6570\u5c5e\u6027\u7684\u5f71\u54cd\u3002\u4f46\u6240\u63d0\u51fa\u7684\u7a0b\u5e8f\u6709\u8bb8\u591a\u5c1a\u672a\u89e3\u51b3\u7684\u95ee\u9898\uff0c\u5728\u62a5\u544a\u7684\u540e\u534a\u90e8\u5206\uff0c\u62a5\u544a\u4eba\u4ecb\u7ecd\u65b0\u7684\u6570\u503c\u7a33\u5b9a\u7684\u7ecf\u9a8c\u9ad8\u9636\u5f71\u54cd\u51fd\u6570\u3002\u65b0\u7684\u9ad8\u9636\u5f71\u54cd\u51fd\u6570\u4e0e\u4e4b\u524d\u63d0\u51fa\u7684\u9ad8\u9636\u5f71\u54cd\u51fd\u6570\u5177\u6709\u975e\u5e38\u76f8\u4f3c\u7684\u6e10\u8fdb\u7279\u6027\uff0c\u4f46\u5374\u4eab\u6709\u66f4\u597d\u7684\u6709\u9650\u6837\u672c\u7279\u6027\u3002\u6700\u540e\uff0c\u901a\u8fc7\u6a21\u62df\u7814\u7a76\u8bc1\u660e\u6539\u8fdb\u7684\u6709\u9650\u6837\u672c\u6027\u80fd\u3002<\/span><\/section>\n

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

Causal mediation analysis with Mendelian Randomization<\/span><\/strong><\/span><\/h4>\n
Jingshu Wang, University of Chicago<\/span><\/section>\n
\u4e86\u89e3\u5e38\u89c1\u75be\u75c5\u7684\u81f4\u75c5\u673a\u5236\u662f\u4e34\u5e8a\u7814\u7a76\u7684\u4e00\u4e2a\u57fa\u672c\u76ee\u6807\u3002\u7531\u4e8e\u968f\u673a\u5bf9\u7167\u5b9e\u9a8c\u5e76\u4e0d\u603b\u662f\u53ef\u884c\u7684\uff0c\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\uff08MR\uff09\uff0c\u4f7f\u7528\u81ea\u7136\u9057\u4f20\u7a81\u53d8\u4f5c\u4e3a\u5de5\u5177\uff0c\u5df2\u7ecf\u6210\u4e3a\u63a2\u7a76\u5e38\u89c1\u75be\u75c5\u56e0\u679c\u673a\u5236\u7684\u4e00\u4e2a\u6d41\u884c\u7684\u66ff\u4ee3\u65b9\u6cd5\u3002\u7136\u800c\uff0c\u76ee\u524d\u7684MR\u65b9\u6cd5\u901a\u5e38\u5ffd\u7565\u4e86\u98ce\u9669\u56e0\u7d20\u548c\u75be\u75c5\u8fdb\u5c55\u4e4b\u95f4\u7684\u65f6\u95f4\u5173\u7cfb\u3002\u5728\u672c\u8bb2\u5ea7\u4e2d\uff0c\u6211\u5c06\u8ba8\u8bba\u4e00\u79cd\u57fa\u4e8e\u5b5f\u5fb7\u5c14\u968f\u673a\u5316\u7684\u7edf\u8ba1\u65b9\u6cd5\uff0c\u4ee5\u8bc4\u4f30\u4e00\u8fde\u4e32\u98ce\u9669\u56e0\u7d20\u5728\u65f6\u95f4\u987a\u5e8f\u4e0a\u5bf9\u540e\u6765\u75be\u75c5\u72b6\u51b5\u7684\u56e0\u679c\u8c03\u89e3\u6548\u5e94\u3002\u4e3a\u4e86\u63d0\u9ad8\u6548\u7387\u548c\u7a33\u5065\u6027\uff0c\u6211\u4eec\u7684\u6846\u67b6\u662f\u57fa\u4e8e\u4e00\u4e2a\u5b8c\u6574\u7684\u8d1d\u53f6\u65af\u6846\u67b6\uff0c\u5e76\u5141\u8bb8\u8c03\u6574\u591a\u6001\u6027\u6548\u5e94\u3002\u6211\u5c06\u5728\u6a21\u62df\u548c\u771f\u5b9e\u6570\u636e\u6848\u4f8b\u7814\u7a76\u4e2d\u8bf4\u660e\u6211\u4eec\u65b9\u6cd5\u7684\u6027\u80fd\u3002<\/span><\/section>\n


<\/h4>\n

Safe Policy Learning through Extrapolation<\/span><\/strong><\/span><\/h4>\n
Zhichao Jiang, Sun Yat-sen Univeristy<\/span><\/section>\n
\u57fa\u4e8e\u7b97\u6cd5\u7684\u5efa\u8bae\u548c\u51b3\u7b56\u5728\u5f53\u4eca\u793e\u4f1a\u5df2\u7ecf\u65e0\u5904\u4e0d\u5728\u4e86\u3002\u8bb8\u591a\u8fd9\u4e9b\u548c\u5176\u4ed6\u6570\u636e\u9a71\u52a8\u7684\u653f\u7b56\u90fd\u662f\u57fa\u4e8e\u5df2\u77e5\u7684\u3001\u786e\u5b9a\u7684\u89c4\u5219\uff0c\u4ee5\u786e\u4fdd\u5176\u900f\u660e\u5ea6\u548c\u53ef\u89e3\u91ca\u6027\u3002\u5f53\u8fd9\u4e9b\u653f\u7b56\u88ab\u7528\u4e8e\u516c\u5171\u653f\u7b56\u51b3\u7b56\u65f6\uff0c\u60c5\u51b5\u5c24\u5176\u5982\u6b64\u3002\u4f8b\u5982\uff0c\u4f5c\u4e3a\u6211\u4eec\u7684\u6fc0\u52b1\u6027\u5e94\u7528\u7684\u5ba1\u524d\u98ce\u9669\u8bc4\u4f30\u7b97\u6cd5\uff0c\u63d0\u4f9b\u4e86\u76f8\u5bf9\u7b80\u5355\u3001\u786e\u5b9a\u7684\u5206\u7c7b\u5206\u6570\u548c\u5efa\u8bae\uff0c\u4ee5\u5e2e\u52a9\u6cd5\u5b98\u505a\u51fa\u91ca\u653e\u51b3\u5b9a\u3002\u4e0d\u5e78\u7684\u662f\uff0c\u73b0\u6709\u7684\u653f\u7b56\u5b66\u4e60\u65b9\u6cd5\u5e76\u4e0d\u9002\u7528\uff0c\u56e0\u4e3a\u5b83\u4eec\u8981\u6c42\u73b0\u6709\u7684\u653f\u7b56\u662f\u968f\u673a\u7684\u800c\u4e0d\u662f\u786e\u5b9a\u6027\u7684\u3002\u6211\u4eec\u5f00\u53d1\u4e86\u4e00\u79cd\u7a33\u5065\u7684\u4f18\u5316\u65b9\u6cd5\uff0c\u5b83\u53ef\u4ee5\u90e8\u5206\u5730\u786e\u5b9a\u653f\u7b56\u7684\u9884\u671f\u6548\u7528\uff0c\u7136\u540e\u901a\u8fc7\u6700\u5c0f\u5316\u6700\u574f\u60c5\u51b5\u4e0b\u7684\u635f\u5931\u627e\u5230\u4e00\u4e2a\u6700\u4f73\u653f\u7b56\u3002\u7531\u6b64\u4ea7\u751f\u7684\u653f\u7b56\u662f\u4fdd\u5b88\u7684\uff0c\u4f46\u6709\u7edf\u8ba1\u5b66\u4e0a\u7684\u5b89\u5168\u4fdd\u8bc1\uff0c\u5141\u8bb8\u653f\u7b56\u5236\u5b9a\u8005\u9650\u5236\u4ea7\u751f\u6bd4\u73b0\u6709\u653f\u7b56\u66f4\u7cdf\u7cd5\u7684\u7ed3\u679c\u7684\u6982\u7387\u3002\u6700\u540e\uff0c\u62a5\u544a\u4eba\u5c06\u63d0\u51fa\u7684\u65b9\u6cd5\u5e94\u7528\u4e8e\u4e00\u4e2a\u72ec\u7279\u7684\u5ba1\u5224\u524d\u98ce\u9669\u8bc4\u4f30\u7684\u73b0\u573a\u5b9e\u9a8c\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Ben-Michael, E., Greiner, D. J., Imai, K., & Jiang, Z. (2021). Safe Policy Learning through Extrapolation: Application to Pre-trial Risk Assessment. arXiv preprint arXiv:2109.11679.<\/span><\/section>\n


<\/h4>\n

Evaluating Causes of Effects by Posterior Effects of Causes<\/span><\/strong><\/span><\/h4>\n
Jinzhu Jia, Peking University<\/span><\/section>\n
\u5bf9\u4e8e\u5355\u4e00\u56e0\u679c\u53d8\u91cf\u7684\u60c5\u51b5\uff0cDawid\u7b49\u4eba\uff082014\uff09\u5b9a\u4e49\u4e86\u56e0\u679c\u6982\u7387\uff0cPearl\uff082000\uff09\u5b9a\u4e49\u4e86\u5fc5\u8981\u6027\u6982\u7387\u6765\u8bc4\u4f30\u6548\u5e94\u7684\u539f\u56e0\u3002\u5bf9\u4e8e\u6709\u591a\u4e2a\u53ef\u80fd\u76f8\u4e92\u5f71\u54cd\u7684\u539f\u56e0\u7684\u60c5\u51b5\uff0c\u672c\u6587\u6839\u636e\u89c2\u5bdf\u5230\u7684\u6cbb\u7597\u540e\u53d8\u91cf\u7684\u8bc1\u636e\uff0c\u5b9a\u4e49\u4e86\u540e\u9a8c\u603b\u56e0\u679c\u6548\u5e94\u548c\u76f4\u63a5\u56e0\u679c\u6548\u5e94\uff0c\u8fd9\u53ef\u4ee5\u88ab\u770b\u4f5c\u662f\u5bf9\u6548\u5e94\u539f\u56e0\u7684\u6d4b\u91cf\u3002\u540e\u53d1\u56e0\u679c\u6548\u5e94\u6d89\u53ca\u53cd\u4e8b\u5b9e\u53d8\u91cf\u7684\u6982\u7387\u3002\u56e0\u6b64\uff0c\u4e0e\u56e0\u679c\u6982\u7387\u3001\u5fc5\u8981\u6027\u7684\u6982\u7387\u548c\u76f4\u63a5\u56e0\u679c\u6548\u5e94\u4e00\u6837\uff0c\u540e\u9a8c\u603b\u56e0\u679c\u6548\u5e94\u548c\u76f4\u63a5\u56e0\u679c\u6548\u5e94\u7684\u53ef\u8bc6\u522b\u6027\u6bd4\u4f20\u7edf\u7684\u4ee5\u6cbb\u7597\u524d\u53d8\u91cf\u4e3a\u6761\u4ef6\u7684\u56e0\u679c\u6548\u5e94\u7684\u53ef\u8bc6\u522b\u6027\u9700\u8981\u66f4\u591a\u7684\u5047\u8bbe\u3002\u6211\u4eec\u63d0\u51fa\u4e86\u540e\u53d1\u56e0\u679c\u6548\u5e94\u7684\u53ef\u8bc6\u522b\u6027\u6240\u9700\u7684\u5047\u8bbe\uff0c\u5e76\u63d0\u4f9b\u4e86\u8bc6\u522b\u65b9\u7a0b\u3002\u6b64\u5916\uff0c\u5f53\u591a\u4e2a\u539f\u56e0\u548c\u4e00\u4e2a\u7ec8\u70b9\u4e4b\u95f4\u7684\u56e0\u679c\u5173\u7cfb\u53ef\u4ee5\u7528\u56e0\u679c\u7f51\u7edc\u6765\u63cf\u8ff0\u65f6\uff0c\u6211\u4eec\u53ef\u4ee5\u7b80\u5316\u6240\u9700\u7684\u5047\u8bbe\u4ee5\u53ca\u540e\u9a8c\u603b\u56e0\u679c\u6548\u5e94\u548c\u76f4\u63a5\u56e0\u679c\u6548\u5e94\u7684\u8bc6\u522b\u65b9\u7a0b\u3002\u6700\u540e\uff0c\u901a\u8fc7\u6570\u5b57\u5b9e\u4f8b\uff0c\u6211\u4eec\u5c06\u540e\u9a8c\u603b\u6548\u5e94\u548c\u76f4\u63a5\u56e0\u679c\u6548\u5e94\u4e0e\u5176\u4ed6\u8bc4\u4f30\u6548\u5e94\u539f\u56e0\u548c\u4eba\u7fa4\u5f52\u56e0\u98ce\u9669\u7684\u63aa\u65bd\u8fdb\u884c\u4e86\u6bd4\u8f83\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Lu, Z., Geng, Z., Li, W., Zhu, S., & Jia, J. (2022). Evaluating Causes of Effects by Posterior Effects of Causes. Biometrika.<\/span><\/section>\n


<\/span><\/h4>\n

Retrospective causal inference with multiple effect variables<\/span><\/strong><\/span><\/h4>\n
Wei Li, Renmin University of China<\/span><\/section>\n
\u6b63\u5982Dawid\uff082000\uff09\u548cPearl & Mackenzie\uff082018\uff09\u6240\u5f3a\u8c03\u7684\uff0c\u5728\u56e0\u679c\u63a8\u65ad\u4e2d\uff0c\u63a8\u5bfc\u7ed9\u5b9a\u6548\u679c\u7684\u539f\u56e0\u662f\u4e00\u4e2a\u6bd4\u8bc4\u4ef7\u539f\u56e0\u7684\u6548\u679c\u66f4\u5177\u6311\u6218\u6027\u7684\u95ee\u9898\u3002Lu\u7b49\u4eba\uff082022\uff09\u63d0\u51fa\u4e86\u4e00\u79cd\u57fa\u4e8e\u540e\u9a8c\u56e0\u679c\u6548\u5e94\u63a8\u5bfc\u5355\u4e00\u6548\u5e94\u53d8\u91cf\u7684\u539f\u56e0\u7684\u65b9\u6cd5\u3002\u5728\u8bb8\u591a\u5e94\u7528\u4e2d\uff0c\u6709\u591a\u4e2a\u6548\u5e94\u53d8\u91cf\uff0c\u56e0\u6b64\u53ef\u4ee5\u540c\u65f6\u4f7f\u7528\u5b83\u4eec\u6765\u66f4\u51c6\u786e\u5730\u63a8\u5bfc\u539f\u56e0\u3002\u4e3a\u4e86\u4ece\u591a\u4e2a\u6548\u5e94\u4e2d\u56de\u987e\u6027\u5730\u63a8\u5bfc\u51fa\u539f\u56e0\uff0c\u6211\u4eec\u63d0\u51fa\u4e86\u4ee5\u89c2\u5bdf\u5230\u7684\u8bc1\u636e\u4e3a\u6761\u4ef6\u7684\u591a\u53d8\u91cf\u540e\u9a8c\u603b\u6548\u5e94\u3001\u5e72\u9884\u6548\u5e94\u548c\u76f4\u63a5\u56e0\u679c\u6548\u5e94\u3002\u6211\u4eec\u63cf\u8ff0\u4e86\u65e0\u6df7\u6742\u548c\u5355\u8c03\u6027\u7684\u5047\u8bbe\uff0c\u5728\u6b64\u5047\u8bbe\u4e0b\uff0c\u6211\u4eec\u8bc1\u660e\u4e86\u591a\u53d8\u91cf\u540e\u9a8c\u56e0\u679c\u6548\u5e94\u7684\u53ef\u8bc6\u522b\u6027\uff0c\u5e76\u63d0\u4f9b\u4e86\u5176\u8bc6\u522b\u65b9\u7a0b\u3002\u5f53\u539f\u56e0\u548c\u7ed3\u679c\u4e4b\u95f4\u7684\u56e0\u679c\u5173\u7cfb\u7531\u56e0\u679c\u7f51\u7edc\u63cf\u8ff0\u65f6\uff0c\u5047\u8bbe\u548c\u8bc6\u522b\u65b9\u7a0b\u90fd\u53ef\u4ee5\u7b80\u5316\u3002\u6240\u63d0\u51fa\u7684\u65b9\u6cd5\u53ef\u4ee5\u5e94\u7528\u4e8e\u6709\u591a\u4e2a\u6548\u5e94\u6216\u7ed3\u679c\u53d8\u91cf\u7684\u5404\u79cd\u7814\u7a76\u4e2d\u7684\u56e0\u679c\u5f52\u5c5e\u3001\u533b\u7597\u8bca\u65ad\u7b49\u3002<\/span><\/section>\n


<\/h2>\n


<\/p>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u516b\uff08Kun Kuang\uff0cChair\uff09<\/strong> <\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Learning Causal Representations for Generalization in Reinforcement Learning<\/span><\/strong><\/span><\/h4>\n
Chaochao Lu, University of Cambridge & Max Planck Institute for Intelligent Systems<\/span><\/section>\n
\u6a21\u4eff\u5b66\u4e60\u548c\u5f3a\u5316\u5b66\u4e60\u7684\u4e00\u4e2a\u57fa\u672c\u6311\u6218\u662f\u5b66\u4e60\u653f\u7b56\u3001\u8868\u5f81\u6216\u52a8\u6001\uff0c\u8fd9\u4e9b\u653f\u7b56\u3001\u8868\u5f81\u6216\u52a8\u6001\u4e0d\u5efa\u7acb\u5728\u865a\u5047\u7684\u76f8\u5173\u5173\u7cfb\u4e0a\uff0c\u5e76\u80fd\u8d85\u8d8a\u5b83\u4eec\u88ab\u8bad\u7ec3\u7684\u7279\u5b9a\u73af\u5883\u3002\u6211\u4eec\u4ece\u4e00\u4e2a\u7edf\u4e00\u7684\u89d2\u5ea6\u6765\u7814\u7a76\u8fd9\u4e9b\u6cdb\u5316\u95ee\u9898\u3002\u4e3a\u6b64\uff0c\u6211\u4eec\u63d0\u51fa\u4e86\u4e00\u4e2a\u901a\u7528\u7684\u6846\u67b6\u6765\u89e3\u51b3\u8fd9\u4e9b\u95ee\u9898\uff0c\u5728\u73af\u5883\u53d8\u5316\u7684\u6e29\u548c\u5047\u8bbe\u4e0b\uff0c\u5728\u7406\u8bba\u4e0a\u4fdd\u8bc1\u53ef\u8bc6\u522b\u6027\u548c\u6cdb\u5316\u6027\u3002\u901a\u8fc7\u5229\u7528\u4e00\u7ec4\u4e0d\u540c\u7684\u8bad\u7ec3\u73af\u5883\uff0c\u6211\u4eec\u6784\u5efa\u4e86\u4e00\u4e2a\u6570\u636e\u8868\u793a\u6cd5\uff0c\u8be5\u8868\u793a\u6cd5\u5ffd\u7565\u4e86\u4efb\u4f55\u865a\u5047\u7684\u7279\u5f81\uff0c\u5e76\u5728\u4e0d\u540c\u7684\u73af\u5883\u4e2d\u4e00\u81f4\u5730\u9884\u6d4b\u76ee\u6807\u53d8\u91cf\u3002\u6309\u7167\u8fd9\u79cd\u65b9\u6cd5\uff0c\u6211\u4eec\u5728\u7b56\u7565\u3001\u8868\u793a\u548c\u52a8\u6001\u65b9\u9762\u5efa\u7acb\u4e86\u4e0d\u53d8\u7684\u9884\u6d4b\u5668\u3002\u6211\u4eec\u4ece\u7406\u8bba\u4e0a\u8868\u660e\uff0c\u6240\u4ea7\u751f\u7684\u653f\u7b56\u3001\u8868\u793a\u548c\u52a8\u6001\u80fd\u591f\u63a8\u5e7f\u5230\u672a\u89c1\u8fc7\u7684\u73af\u5883\u3002\u5728\u5408\u6210\u548c\u771f\u5b9e\u4e16\u754c\u7684\u6570\u636e\u96c6\u4e0a\u8fdb\u884c\u7684\u5e7f\u6cdb\u5b9e\u9a8c\u8868\u660e\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u6bd4\u5404\u79cd\u57fa\u7ebf\u8fbe\u5230\u4e86\u66f4\u597d\u7684\u6cdb\u5316\u6548\u679c\u3002<\/span><\/section>\n

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

Causal Inference with Instrumental Variables<\/span><\/strong><\/span><\/h4>\n
Kun Kuang, Zhejiang University<\/span><\/section>\n
\u56e0\u679c\u95ee\u9898\u5b58\u5728\u4e8e\u8bb8\u591a\u9886\u57df\uff0c\u5982\u533b\u7597\u4fdd\u5065\u3001\u7ecf\u6d4e\u5b66\u3001\u653f\u6cbb\u5b66\u3001\u6570\u5b57\u8425\u9500\u7b49\u3002\u4e0e\u65e7\u7684\u836f\u7269\u76f8\u6bd4\uff0c\u4e00\u79cd\u65b0\u7684\u836f\u7269\u662f\u5426\u80fd\u4f7f\u67d0\u79cd\u75be\u75c5\u7684\u6cbb\u7597\u6548\u679c\u66f4\u597d\uff1f\u4e00\u79cd\u65b0\u7684\u8425\u9500\u7b56\u7565\u662f\u5426\u80fd\u63d0\u9ad8\u67d0\u79cd\u4ea7\u54c1\u7684\u9500\u91cf\uff1f\u6240\u6709\u8fd9\u4e9b\u95ee\u9898\u90fd\u53ef\u4ee5\u901a\u8fc7\u56e0\u679c\u63a8\u7406\u6280\u672f\u6765\u89e3\u51b3\u3002\u56e0\u679c\u63a8\u65ad\u7684\u91d1\u6807\u51c6\u65b9\u6cd5\u662f\u968f\u673a\u5b9e\u9a8c\uff0c\u4f8b\u5982A\/B\u6d4b\u8bd5\u3002\u7136\u800c\uff0c\u5b8c\u5168\u968f\u673a\u5316\u7684\u5b9e\u9a8c\u901a\u5e38\u975e\u5e38\u6602\u8d35\uff0c\u6709\u65f6\u751a\u81f3\u4e0d\u53ef\u884c\u3002\u56e0\u6b64\uff0c\u5f00\u53d1\u81ea\u52a8\u7edf\u8ba1\u65b9\u6cd5\u6765\u63a8\u65ad\u89c2\u5bdf\u6027\u7814\u7a76\u4e2d\u7684\u56e0\u679c\u6548\u5e94\u662f\u975e\u5e38\u56f0\u96be\u7684\u3002<\/span><\/section>\n
\u672c\u6b21\u62a5\u544a\u5c55\u793a\u4e86\u5927\u6570\u636e\u573a\u666f\u4e0b\u56e0\u679c\u63a8\u65ad\u7684\u4e00\u4e9b\u65b0\u6311\u6218\uff0c\u5305\u62ec\uff081\uff09\u9ad8\u7ef4\u548c\u566a\u58f0\u53d8\u91cf\uff0c\uff082\uff09\u672a\u89c2\u5bdf\u5230\u7684\u6df7\u6742\u56e0\u7d20\uff0c\u4ee5\u53ca\uff083\uff09\u590d\u6742\u7684\u6cbb\u7597\u53d8\u91cf\u3002\u6211\u4eec\u5c06\u4e3b\u8981\u5173\u6ce8\u6765\u81ea\u672a\u89c2\u5bdf\u5230\u7684\u6df7\u6742\u56e0\u7d20\u7684\u6311\u6218\uff0c\u5e76\u4ecb\u7ecd\u6700\u8fd1\u5728\u56e0\u679c\u63a8\u65ad\u548c\u673a\u5668\u5b66\u4e60\u754c\u63d0\u51fa\u7684IV\u65b9\u6cd5\u3002\u5177\u4f53\u6765\u8bf4\uff0c\u6211\u4eec\u9996\u5148\u4ecb\u7ecd\u5982\u4f55\u7ed3\u5408\u6df7\u6742\u56e0\u7d20\u5e73\u8861\u6280\u672f\u548cIV\u56de\u5f52\u6a21\u578b\u6765\u5b9e\u73b0\u6df7\u6742\u56e0\u7d20\u5e73\u8861IV\u56de\u5f52\u3002\u7136\u540e\uff0c\u6211\u4eec\u5c06\u8ba8\u8bba\u5982\u4f55\u4ece\u89c2\u5bdf\u5230\u7684\u53d8\u91cf\u4e2d\u751f\u6210\u4e00\u4e2a\u8868\u5f81\u6765\u670d\u52a1\u4e8eIVs\u7684\u4f5c\u7528\u3002\u6700\u540e\uff0c\u6211\u4eec\u5c06\u4ecb\u7ecd\u5982\u4f55\u7528Meta-EM\u7b97\u6cd5\u4ece\u6570\u636e\u878d\u5408\u4e2d\u5b66\u4e60\u6f5c\u5728\u7684\u7fa4\u4f53IV\uff0c\u4ee5\u8fdb\u884c\u5b58\u5728\u672a\u89c2\u5bdf\u5230\u7684\u6df7\u6742\u56e0\u7d20\u7684\u56e0\u679c\u63a8\u65ad\u3002<\/span><\/section>\n


<\/h2>\n


<\/p>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u4e5d\uff08Kun Zhang\uff0cChair\uff09<\/strong><\/span><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Causal Inference of Truncation-by-Death with Unmeasured Confounding<\/span><\/strong><\/span><\/h4>\n
Xiao-Hua Zhou, Peking University<\/span><\/section>\n
\u4e34\u5e8a\u7814\u7a76\u7ecf\u5e38\u9047\u5230\u6b7b\u4ea1\u622a\u65ad\u7684\u60c5\u51b5\uff0c\u8fd9\u53ef\u80fd\u4f7f\u4e00\u4e9b\u7ed3\u679c\u65e0\u6cd5\u786e\u5b9a\u3002\u4ec5\u4ec5\u57fa\u4e8e\u89c2\u5bdf\u5230\u7684\u5e78\u5b58\u8005\u7684\u7edf\u8ba1\u5206\u6790\u53ef\u80fd\u4f1a\u5bfc\u81f4\u6709\u504f\u5dee\u7684\u7ed3\u679c\uff0c\u56e0\u4e3a\u4e0d\u540c\u6cbb\u7597\u7ec4\u7684\u5e78\u5b58\u8005\u7684\u7279\u5f81\u53ef\u80fd\u4e0d\u540c\u3002\u5728\u8fd9\u79cd\u60c5\u51b5\u4e0b\uff0c\u5e38\u7528\u7684\u6709\u610f\u4e49\u7684\u56e0\u679c\u53c2\u6570\u662f\u5e78\u5b58\u8005\u5e73\u5747\u56e0\u679c\u6548\u5e94\uff08SACE\uff09\uff0c\u5f53\u6cbb\u7597\u5206\u914d\u548c\u751f\u5b58\u6216\u7ed3\u679c\u8fc7\u7a0b\u4e4b\u95f4\u5b58\u5728\u65e0\u6cd5\u6d4b\u91cf\u7684\u6df7\u6742\u65f6\uff0c\u8be5\u53c2\u6570\u53ef\u80fd\u65e0\u6cd5\u8bc6\u522b\u3002\u5728\u672c\u62a5\u544a\u4e2d\uff0c\u6211\u4eec\u9996\u5148\u8868\u660e\uff0c\u5728\u9002\u5f53\u7684\u5047\u8bbe\u6761\u4ef6\u4e0b\uff0c\u57fa\u4e8e\u66ff\u4ee3\u53d8\u91cf\u7684\u5e78\u5b58\u8005\u5e73\u5747\u56e0\u679c\u6548\u5e94\u662f\u53ef\u4ee5\u8bc6\u522b\u7684\u3002\u63a5\u4e0b\u6765\uff0c\u6211\u4eec\u4e3a\u8fd9\u4e2a\u4f30\u8ba1\u503c\u63d0\u51fa\u4e86\u4e00\u4e2a\u589e\u5f3a\u7684\u9006\u6982\u7387\u52a0\u6743\uff08AIPW\uff09\u7c7b\u578b\u7684\u4f30\u8ba1\u5668\uff0c\u5bf9\u6a21\u578b\u7684\u9519\u8bef\u6307\u5b9a\u5177\u6709\u9c81\u68d2\u6027\u3002\u6700\u540e\uff0c\u5efa\u8bae\u7684\u65b9\u6cd5\u88ab\u5e94\u7528\u4e8e\u7814\u7a76\u5f02\u4f53\u5e72\u7ec6\u80de\u79fb\u690d\u7c7b\u578b\u5bf9\u767d\u8840\u75c5\u590d\u53d1\u7684\u5f71\u54cd\u3002<\/span><\/section>\n

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

Causality-inspired ML: what can causality do for ML?<\/span><\/strong><\/span><\/h4>\n
Sara Magliacane, University of Amsterdam & Research Scientist at IBM Watson AI Lab<\/span><\/section>\n
\u5c06\u673a\u5668\u5b66\u4e60\u5e94\u7528\u5230\u73b0\u5b9e\u4e16\u754c\u7684\u6848\u4f8b\u4e2d\uff0c\u5f80\u5f80\u9700\u8981\u5bf9\u5f02\u8d28\u6027\u3001\u975e\u968f\u673a\u7f3a\u5931\u6216\u635f\u574f\u7684\u6570\u636e\u3001\u9009\u62e9\u504f\u5dee\u3001\u975ei.i.d.\u6570\u636e\u7b49\u5177\u6709\u9c81\u68d2\u6027\u7684\u65b9\u6cd5\uff0c\u5e76\u4e14\u80fd\u591f\u5728\u4e0d\u540c\u9886\u57df\u4e2d\u8fdb\u884c\u63a8\u5e7f\u3002\u6b64\u5916\uff0c\u8bb8\u591a\u4efb\u52a1\u672c\u8d28\u4e0a\u662f\u8bd5\u56fe\u56de\u7b54\u56e0\u679c\u95ee\u9898\u5e76\u6536\u96c6\u53ef\u64cd\u4f5c\u7684\u89c1\u89e3\uff0c\u5bf9\u4e8e\u8fd9\u9879\u4efb\u52a1\u6765\u8bf4\uff0c\u76f8\u5173\u5173\u7cfb\u901a\u5e38\u662f\u4e0d\u591f\u7684\u3002\u5728\u4e30\u5bcc\u7684\u56e0\u679c\u63a8\u7406\u6587\u732e\u4e2d\uff0c\u8fd9\u4e9b\u95ee\u9898\u4e2d\u6709\u51e0\u4e2a\u5f97\u5230\u4e86\u89e3\u51b3\u3002\u53e6\u4e00\u65b9\u9762\uff0c\u901a\u5e38\u7ecf\u5178\u7684\u56e0\u679c\u63a8\u65ad\u65b9\u6cd5\u9700\u8981\u5bf9\u56e0\u679c\u56fe\u6709\u5b8c\u6574\u7684\u4e86\u89e3\uff0c\u6216\u8005\u9700\u8981\u6709\u8db3\u591f\u7684\u5b9e\u9a8c\u6570\u636e\uff08\u5e72\u9884\uff09\u6765\u51c6\u786e\u4f30\u8ba1\u5b83\u3002\u6700\u8fd1\uff0c\u4e00\u4e2a\u65b0\u7684\u7814\u7a76\u65b9\u5411\u96c6\u4e2d\u5728\u56e0\u679c\u5173\u7cfb\u542f\u53d1\u7684\u673a\u5668\u5b66\u4e60\u4e0a\uff0c\u5373\u628a\u56e0\u679c\u63a8\u7406\u7684\u601d\u60f3\u5e94\u7528\u5230\u673a\u5668\u5b66\u4e60\u65b9\u6cd5\u4e0a\uff0c\u800c\u4e0d\u4e00\u5b9a\u77e5\u9053\u6216\u751a\u81f3\u8bd5\u56fe\u4f30\u8ba1\u5b8c\u6574\u7684\u56e0\u679c\u56fe\u3002\u5728\u8fd9\u6b21\u6f14\u8bb2\u4e2d\uff0c\u6211\u5c06\u4ecb\u7ecd\u5728\u65e0\u76d1\u7763\u9886\u57df\u9002\u5e94\u60c5\u51b5\u4e0b\u8fd9\u4e00\u7814\u7a76\u65b9\u5411\u7684\u4e00\u4e2a\u4f8b\u5b50\uff0c\u5728\u8fd9\u79cd\u60c5\u51b5\u4e0b\uff0c\u6211\u4eec\u6709\u4e00\u7ec4\u6e90\u9886\u57df\u7684\u6807\u7b7e\u6570\u636e\u548c\u76ee\u6807\u9886\u57df\uff08”zero-shot”\uff09\u7684\u65e0\u6807\u7b7e\u6570\u636e\uff0c\u6211\u4eec\u60f3\u9884\u6d4b\u8fd9\u4e9b\u6807\u7b7e\u3002\u7279\u522b\u662f\uff0c\u8003\u8651\u5230\u67d0\u4e9b\u5047\u8bbe\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u80fd\u591f\u9009\u62e9\u4e00\u7ec4\u88ab\u8bc1\u660e\u662f “\u7a33\u5b9a\u7684”\u7279\u5f81\uff08\u5206\u79bb\u96c6\uff09\uff0c\u5bf9\u4e8e\u8fd9\u4e9b\u7279\u5f81\uff0c\u5373\u4f7f\u5728\u4efb\u610f\u5927\u7684\u5206\u5e03\u53d8\u5316\u7684\u60c5\u51b5\u4e0b\uff0c\u6cdb\u5316\u8bef\u5dee\u4e5f\u53ef\u4ee5\u88ab\u7ea6\u675f\u3002\u4e0e\u5176\u4ed6\u5de5\u4f5c\u4e0d\u540c\u7684\u662f\uff0c\u5b83\u8fd8\u5229\u7528\u4e86\u672a\u6807\u8bb0\u7684\u76ee\u6807\u6570\u636e\u4e2d\u7684\u4fe1\u606f\uff0c\u5141\u8bb8\u5728\u6e90\u57df\u4e2d\u51fa\u73b0\u4e00\u4e9b\u672a\u89c1\u8fc7\u7684\u53d8\u5316\u3002\u867d\u7136\u4f7f\u7528\u4e86\u56e0\u679c\u63a8\u7406\u7684\u601d\u60f3\uff0c\u4f46\u6211\u4eec\u7684\u65b9\u6cd5\u4ece\u6765\u4e0d\u4ee5\u91cd\u5efa\u56e0\u679c\u56fe\u6216\u751a\u81f3\u9a6c\u5c14\u79d1\u592b\u7b49\u4ef7\u7c7b\u4e3a\u76ee\u7684\uff0c\u8fd9\u8868\u660e\u56e0\u679c\u63a8\u7406\u7684\u601d\u60f3\u751a\u81f3\u5728\u8fd9\u79cd\u66f4\u5bbd\u677e\u7684\u73af\u5883\u4e2d\u4e5f\u80fd\u5e2e\u52a9\u673a\u5668\u5b66\u4e60\u3002<\/span><\/section>\n

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

Efficient learning using privileged information with known causal structure<\/span><\/strong><\/span><\/h4>\n
Fredrik Johansson, Chalmers University of Technology, Sweden<\/span><\/section>\n
\u5728\u6837\u672c\u91cf\u6709\u9650\u7684\u9886\u57df\uff0c\u9ad8\u6548\u7684\u5b66\u4e60\u662f\u81f3\u5173\u91cd\u8981\u7684\u3002\u7136\u800c\uff0c\u5bf9\u4e8e\u8bb8\u591a\u673a\u5668\u5b66\u4e60\u95ee\u9898\uff0c\u6807\u51c6\u505a\u6cd5\u901a\u5e38\u4f1a\u7559\u4e0b\u5927\u91cf\u672a\u4f7f\u7528\u7684\u4fe1\u606f\u3002\u4e00\u4e2a\u4f8b\u5b50\u662f\u6839\u636e\u57fa\u7ebf\u65f6\u95f4\u70b9\u6536\u96c6\u7684\u53d8\u91cf\u9884\u6d4b\u4e00\u4e2a\u65f6\u95f4\u5e8f\u5217\u7ed3\u675f\u65f6\u7684\u7ed3\u679c\uff0c\u4f8b\u5982\uff0c\u4e00\u4e2a\u75c5\u4eba\u5165\u9662\u540e30\u5929\u7684\u6b7b\u4ea1\u98ce\u9669\u3002\u5728\u5e94\u7528\u4e2d\uff0c\u5728\u57fa\u7ebf\u548c\u7ec8\u70b9\u4e4b\u95f4\u6536\u96c6\u7684\u4e2d\u95f4\u6837\u672c\u88ab\u4e22\u5f03\u662f\u5f88\u5e38\u89c1\u7684\uff0c\u56e0\u4e3a\u5728\u4f7f\u7528\u5b66\u4e60\u6a21\u578b\u65f6\uff0c\u5b83\u4eec\u4e0d\u80fd\u4f5c\u4e3a\u9884\u6d4b\u7684\u8f93\u5165\u3002\u6211\u4eec\u8bf4\uff0c\u8fd9\u4e9b\u4fe1\u606f\u662f\u6709\u7279\u6743\u7684\uff0c\u56e0\u4e3a\u5b83\u53ea\u5728\u8bad\u7ec3\u65f6\u53ef\u7528\u3002\u5728\u8fd9\u6b21\u6f14\u8bb2\u4e2d\uff0c\u6211\u4eec\u8868\u660e\uff0c\u5229\u7528\u5df2\u77e5\u7684\u56e0\u679c\u7ed3\u6784\u548c\u4e2d\u95f4\u65f6\u95f4\u5e8f\u5217\u7684\u7279\u6743\u4fe1\u606f\u53ef\u4ee5\u5bfc\u81f4\u66f4\u6709\u6548\u7684\u5b66\u4e60\u3002\u6211\u4eec\u7ed9\u51fa\u4e86\u8bc1\u660e\u5176\u4f18\u4e8e\u7ecf\u5178\u5b66\u4e60\u7684\u6761\u4ef6\uff0c\u4ee5\u53ca\u652f\u6301\u8fd9\u4e9b\u53d1\u73b0\u7684\u4e00\u7cfb\u5217\u7ecf\u9a8c\u7ed3\u679c\u3002<\/span><\/section>\n

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

Advances in causal representation learning<\/span><\/strong><\/span><\/h4>\n
Kun Zhang, Carnegie Mellon University & Mohamed bin Zayed University of Artificial Intelligence<\/span><\/section>\n
\u672c\u8bb2\u5ea7\u5173\u6ce8\u7684\u662f\u56e0\u679c\u8868\u5f81\u5b66\u4e60\uff0c\u5176\u76ee\u7684\u662f\u63ed\u793a\u5e95\u5c42\u9690\u85cf\u7684\u56e0\u679c\u53d8\u91cf\u53ca\u5176\u5173\u7cfb\u3002\u5b83\u53ef\u4ee5\u88ab\u770b\u4f5c\u662f\u56e0\u679c\u53d1\u73b0\u7684\u4e00\u4e2a\u7279\u4f8b\uff0c\u5176\u76ee\u6807\u662f\u4ece\u89c2\u5bdf\u6570\u636e\u4e2d\u6062\u590d\u6f5c\u5728\u7684\u56e0\u679c\u7ed3\u6784\u6216\u56e0\u679c\u6a21\u578b\u3002\u56e0\u679c\u7cfb\u7edf\u7684\u6a21\u5757\u5316\u5c5e\u6027\u610f\u5473\u7740\u56e0\u679c\u8868\u5f81\u7684\u6700\u5c0f\u53d8\u5316\u548c\u72ec\u7acb\u53d8\u5316\u7684\u5c5e\u6027\uff0c\u6211\u5c06\u89e3\u91ca\u8fd9\u4e9b\u5c5e\u6027\u5982\u4f55\u4f7f\u5f97\u4ece\u89c2\u5bdf\u6570\u636e\u4e2d\u6062\u590d\u5e95\u5c42\u56e0\u679c\u8868\u5f81\u6210\u4e3a\u53ef\u80fd\uff0c\u5e76\u5177\u6709\u53ef\u8bc6\u522b\u6027\u7684\u4fdd\u8bc1\uff1a\u5728\u9002\u5f53\u7684\u5047\u8bbe\u4e0b\uff0c\u5b66\u5230\u7684\u8868\u5f81\u4e0e\u5e95\u5c42\u56e0\u679c\u8fc7\u7a0b\u4e00\u81f4\u3002\u8be5\u8bb2\u5ea7\u5c06\u8003\u8651\u4ee5\u72ec\u7acb\u540c\u5206\u5e03\uff08i.i.d.\uff09\u6570\u636e\u3001\u65f6\u95f4\u6570\u636e\u6216\u6709\u5206\u5e03\u53d8\u5316\u7684\u6570\u636e\u4f5c\u4e3a\u8f93\u5165\u7684\u5404\u79cd\u8bbe\u7f6e\uff0c\u5e76\u8bc1\u660e\u53ef\u8bc6\u522b\u7684\u56e0\u679c\u8868\u5f81\u5b66\u4e60\u4f55\u65f6\u80fd\u4ece\u6df1\u5ea6\u5b66\u4e60\u7684\u7075\u6d3b\u6027\u4e2d\u83b7\u76ca\uff0c\u4f55\u65f6\u5fc5\u987b\u5bf9\u56e0\u679c\u8fc7\u7a0b\u65bd\u52a0\u53c2\u6570\u5316\u7684\u5047\u8bbe\u3002<\/span><\/section>\n

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


<\/h2>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u5341\uff08Thomas Richardson\uff0cChair\uff09<\/strong><\/span><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

On prediction, action and interference in algorithmic fairness<\/span><\/strong><\/span><\/h4>\n
Ricardo Silva, University College London<\/span><\/section>\n
\u6700\u7ec8\uff0c\u6211\u4eec\u5e0c\u671b\u901a\u8fc7\u6539\u53d8\u4e16\u754c\u6765\u51cf\u5c11\u4e0d\u516c\u5e73\u3002\u4ec5\u4ec5\u505a\u51fa\u516c\u6b63\u7684\u88ab\u52a8\u9884\u6d4b\u662f\u4e0d\u591f\u7684\uff0c\u56e0\u6b64\u6211\u4eec\u7684\u51b3\u5b9a\u6700\u7ec8\u5c06\u5bf9\u793e\u4f1a\u7cfb\u7edf\u7684\u8fd0\u4f5c\u4ea7\u751f\u5f71\u54cd\u3002\u6211\u4eec\u5c06\u8ba8\u8bba\u6a21\u62df\u5047\u8bbe\u5e72\u9884\u7684\u65b9\u6cd5\uff0c\u4ee5\u4fbf\u5c0a\u91cd\u53cd\u4e8b\u5b9e\u516c\u5e73\u6027\u7684\u5177\u4f53\u63aa\u65bd\uff1a\u5373\uff0c\u654f\u611f\u6027\u5c5e\u6027\u5982\u4f55\u4e0e\u6211\u4eec\u7684\u884c\u52a8\u76f8\u4e92\u4f5c\u7528\uff0c\u4ece\u800c\u5bfc\u81f4\u4e0d\u516c\u5e73\u7684\u5206\u914d\u7ed3\u679c\uff0c\u5728\u8fd9\u79cd\u60c5\u51b5\u4e0b\uff0c\u6211\u4eec\u5982\u4f55\u5728\u53ef\u884c\u7684\u884c\u52a8\u7a7a\u95f4\u5185\u51cf\u8f7b\u8fd9\u79cd\u4e0d\u5e73\u8861\u7684\u5f71\u54cd\uff1f\u66f4\u4e3a\u56f0\u96be\u7684\u662f\uff0c\u5e72\u6270\u5f88\u53ef\u80fd\u53d1\u751f\uff1a\u4e00\u4e2a\u4eba\u53d1\u751f\u7684\u4e8b\u60c5\u53ef\u80fd\u4f1a\u5f71\u54cd\u53e6\u4e00\u4e2a\u4eba\u3002\u6211\u4eec\u5c06\u8ba8\u8bba\u5982\u4f55\u8868\u8fbe\u8fd9\u4e9b\u56e0\u679c\u56e0\u7d20\u5bf9\u516c\u5e73\u51b3\u7b56\u7684\u5047\u8bbe\u548c\u540e\u679c\uff0c\u627f\u8ba4\u8fd9\u662f\u4e00\u9879\u8270\u5de8\u7684\u4efb\u52a1\uff0c\u4f46\u6211\u4eec\u5e94\u8be5\u5411\u516c\u4f17\u89e3\u91ca\u6211\u4eec\u7684\u63a8\u7406\u3002\u5728\u8ba8\u8bba\u7684\u6700\u540e\uff0c\u6211\u4eec\u5f97\u51fa\u4ee5\u4e0b\u7ed3\u8bba\uff1a\u516c\u5e73\u6700\u7ec8\u662f\u5173\u4e8e\u6539\u53d8\u6570\u636e\u751f\u6210\u8fc7\u7a0b\uff1b\u901a\u8fc7\u4fee\u6539\u9884\u6d4b\u6a21\u578b\u7684\u95f4\u63a5\u53d8\u5316\u662f\u6709\u7528\u7684\uff0c\u4f46\u662f\u7406\u89e3\u5b83\u4eec\u7684\u957f\u671f\u540e\u679c\u53ef\u80fd\u592a\u96be\u4e86\uff1b\u89e3\u51b3\u5e72\u9884\u8fd9\u4e00\u8fdb\u7a0b\u7684\u95ee\u9898\u53ef\u4ee5\u5e26\u6765\u66f4\u591a\u5373\u65f6\u548c\u53ef\u9884\u6d4b\u7684\u7ed3\u679c\uff0c\u5373\u4f7f\u8fd9\u4ecd\u7136\u662f\u4e00\u4e2a\u96be\u9898\uff1b\u53cd\u4e8b\u5b9e\u7279\u6743\u662f\u91cf\u5316\u4e0d\u516c\u5e73\u7684\u4e00\u79cd\u65b9\u6cd5\u3002\u5bf9\u4e8e\u4ece\u4e1a\u4eba\u5458\u800c\u8a00\uff0c\u9700\u8981\u5728\u4f30\u8ba1\u4e2d\u5229\u7528\u4e0d\u786e\u5b9a\u6027\u4ee5\u53ca\u8003\u8651\u8fdd\u53cd\u56e0\u679c\u5047\u8bbe\u7684\u7a33\u5065\u6027\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Kusner, M., Russell, C., Loftus, J., & Silva, R. (2019, May). Making decisions that reduce discriminatory impacts. In International Conference on Machine Learning (pp. 3591-3600). PMLR.<\/span><\/section>\n


<\/h4>\n

Problems due to Selection and an Overcommitment to Ignorability in Causal Fairness<\/span><\/strong><\/span><\/h4>\n
Jake Fawkes, University of Oxford<\/span><\/section>\n
\u5728\u8fd9\u6b21\u62a5\u544a\u4e2d\uff0c\u6211\u4eec\u8ba8\u8bba\u4e86\u54ea\u4e9b\u56e0\u679c\u6a21\u578b\u53ef\u4ee5\u6b63\u786e\u5730\u6355\u6349\u516c\u5e73\u95ee\u9898\u4e2d\u7684\u56e0\u679c\u7279\u5f81\u3002\u5f53\u524d\u7684\u65b9\u6cd5\u6355\u6349\u4e86\u4e00\u4e2a\u76f4\u89c2\u7684\u6982\u5ff5\uff0c\u5373\u5982\u679c\u9884\u6d4b\u4e0e\u5982\u679c\u67d0\u4eba\u7684\u79cd\u65cf\u3001\u6027\u522b\u6216\u5b97\u6559\u4e8b\u5b9e\u4e0a\u76f8\u53cd\uff0c\u90a3\u4e48\u9884\u6d4b\u662f\u516c\u5e73\u7684\u3002\u6211\u4eec\u9996\u5148\u6307\u51fa\uff0c\u5f53\u524d\u7684\u56e0\u679c\u516c\u5e73\u6587\u732e\u901a\u5e38\u81f4\u529b\u4e8e\u5177\u6709\u72ec\u7acb\u566a\u58f0\u548c\u7956\u5148\u5c01\u95ed\u654f\u611f\u5c5e\u6027\u7684DAG\u3002\u56e0\u6b64\uff0c\u6211\u4eec\u8ba4\u4e3a\u8fd9\u4e9b\u5047\u8bbe\u5f80\u5f80\u8fc7\u4e8e\u5f3a\u70c8\uff0c\u65e0\u6cd5\u6355\u6349\u5230\u6b63\u786e\u7684\u56e0\u679c\u7279\u5f81\uff0c\u5982\u53cd\u4e8b\u5b9e\u548c\u56e0\u679c\u6548\u5e94\u3002\u901a\u5e38\u8fd9\u4e9b\u5047\u8bbe\u4ec5\u5728\u968f\u673a\u5bf9\u7167\u8bd5\u9a8c\u4e2d\u51fa\u73b0\uff0c\u6211\u4eec\u8ba4\u4e3a\uff0c\u4e00\u822c\u6765\u8bf4\uff0c\u8fd9\u4e0d\u592a\u53ef\u80fd\u6210\u7acb\uff0c\u8fd9\u7ed9\u53cd\u4e8b\u5b9e\u516c\u5e73\u4ee5\u53ca\u66f4\u4e00\u822c\u7684\u56e0\u679c\u516c\u5e73\u65b9\u6cd5\u7684\u5e94\u7528\u5e26\u6765\u4e86\u56f0\u96be\u3002\u6211\u4eec\u7684\u8bba\u70b9\u57fa\u4e8e\u4e24\u70b9\uff0c\u5373\u8fd9\u4e9b\u6a21\u578b\u81f4\u529b\u4e8e\u53ef\u5ffd\u7565\u6027\u548c\u516c\u5e73\u6027\u6570\u636e\u96c6\u4ea7\u751f\u4e8e\u590d\u6742\u7684\u9009\u62e9\u8fc7\u7a0b\u3002\u6211\u4eec\u63a8\u5bfc\u4e86\u53ef\u5ffd\u7565\u6027\u5fc5\u987b\u6ee1\u8db3\u7684\u6761\u4ef6\uff0c\u5e76\u8ba8\u8bba\u4e86\u5f53\u8be5\u5047\u8bbe\u4e0d\u6210\u7acb\u65f6\u56e0\u679c\u516c\u5e73\u6027\u7684\u542b\u4e49\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Fawkes, J., Evans, R., & Sejdinovic, D. (2022). Selection, Ignorability and Challenges With Causal Fairness. arXiv preprint arXiv:2202.13774.<\/span><\/section>\n


<\/h4>\n

A causal and counterfactual view of (un)fairness in automated decision making<\/span><\/strong><\/span><\/h4>\n
Razieh Nabi, Emory University<\/span><\/section>\n
\u5c3d\u7ba1\u5b58\u5728\u5ba2\u89c2\u6027\u9519\u89c9\uff0c\u4f46\u7b97\u6cd5\u5728\u5176\u53d1\u5c55\u7684\u6bcf\u4e00\u6b65\u90fd\u5229\u7528\u4e86\u4eba\u7c7b\u7684\u4e3b\u89c2\u5224\u65ad\u3002\u5728\u81ea\u52a8\u5316\u51b3\u7b56\u7684\u80cc\u666f\u4e0b\uff0c\u4e00\u4e2a\u7279\u522b\u4ee4\u4eba\u62c5\u5fe7\u7684\u95ee\u9898\u662f\u4f7f\u4e0d\u516c\u6b63\u6c38\u4e45\u5316\uff0c\u5373\u5f53\u6700\u5927\u5316\u201c\u6548\u7528\u201d\u7ef4\u6301\u3001\u5f3a\u5316\u751a\u81f3\u5f15\u5165\u654f\u611f\u7279\u5f81\uff08\u5982\u79cd\u65cf\u3001\u6027\u522b\u3001\u5e74\u9f84\u3001\u6027\u53d6\u5411\uff09\u3001\u51b3\u7b56\u548c\u7ed3\u679c\u4e4b\u95f4\u7684\u4e0d\u516c\u5e73\u4f9d\u8d56\u65f6\u3002\u56e0\u6b64\uff0c\u81ea\u52a8\u5316\u51b3\u7b56\u5fc5\u987b\u5c0a\u91cd\u516c\u5e73\u539f\u5219\uff0c\u7279\u522b\u662f\u5728\u533b\u7597\u3001\u793e\u4f1a\u798f\u5229\u548c\u5211\u4e8b\u53f8\u6cd5\u7b49\u5177\u6709\u793e\u4f1a\u5f71\u54cd\u529b\u7684\u73af\u5883\u4e2d\u3002\u5728\u672c\u6b21\u6f14\u8bb2\u4e2d\uff0c\u6211\u4eec\u5c55\u793a\u4e86\u5982\u4f55\u4f7f\u7528\u56e0\u679c\u63a8\u7406\u548c\u7ea6\u675f\u4f18\u5316\u7684\u65b9\u6cd5\uff0c\u901a\u8fc7\u7ea0\u6b63\u51b3\u7b56\u548c\u7ed3\u679c\u5bf9\u654f\u611f\u7279\u5f81\u7684\u4e0d\u516c\u5e73\u4f9d\u8d56\uff0c\u505a\u51fa\u6700\u4f18\u4f46\u516c\u5e73\u7684\u51b3\u7b56\uff0c\u4ece\u800c\u201c\u6253\u7834\u4e0d\u516c\u5e73\u7684\u5faa\u73af\u201d\u3002\u6211\u4eec\u4ece\u4e09\u4e2a\u65b9\u9762\u5faa\u5e8f\u6e10\u8fdb\u5730\u89e3\u51b3\u81ea\u52a8\u5316\u51b3\u7b56\u4e2d\u7684\u516c\u5e73\u95ee\u9898\uff1a\u9996\u5148\uff0c\u5982\u4f55\u7528\u6570\u5b66\u8868\u8fbe\u516c\u5e73\u539f\u5219\uff1f\u5176\u6b21\uff0c\u5982\u4f55\u4fee\u6539\u7edf\u8ba1\u7a0b\u5e8f\u4ee5\u51cf\u5c11\u4e0d\u516c\u5e73\u5f71\u54cd\uff1f\u4ee5\u53ca\u6700\u540e\uff0c\u5982\u4f55\u63a8\u5e7f\u548c\u90e8\u7f72\u8fd9\u4e9b\u6539\u8fdb\u7b97\u6cd5\uff1f\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u6211\u4eec\u91c7\u53d6\u7684\u65b9\u6cd5\u9700\u8981\u6765\u81ea\u4e13\u5bb6\u548c\/\u6216\u516c\u4f17\u7684\u5927\u91cf\u7684\u4f26\u7406\u8f93\u5165\uff0c\u8fd8\u9700\u8981\u6307\u5b9a\u57fa\u4e8e\u9886\u57df\u77e5\u8bc6\u6216\u8005\u56e0\u679c\u7ed3\u6784\u5b66\u4e60\u540e\u7684\u56e0\u679c\u6a21\u578b\uff0c\u53e6\u5916\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u53ef\u4ee5\u5904\u7406\u672a\u786e\u5b9a\u7684\u56e0\u679c\u5f71\u54cd\uff1b\u5728\u4fee\u6539\u7edf\u8ba1\u7a0b\u5e8f\u65f6\uff0c\u53ef\u4ee5\u4f7f\u7528\u7ea6\u675fMLE\uff08\u6df7\u5408\u4f3c\u7136\uff09\u6216\u8005\u5f00\u53d1\u66f4\u7a33\u5065\u7684\u7ea6\u675f\u4f18\u5316\u65b9\u6cd5\u4ee5\u4f9b\u4f7f\u7528\u5c3d\u53ef\u80fd\u9ad8\u6548\u5730\u6536\u96c6\u6570\u636e\uff1b\u5728\u63a8\u5e7f\u548c\u90e8\u7f72\u7b97\u6cd5\u65f6\uff0c\u9700\u8981\u6ce8\u610f\u6837\u672c\u662f\u5728p\u800c\u975ep*\u4e2d\u91c7\u96c6\u7684\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Nabi, R., Malinsky, D., & Shpitser, I. (2022, June). Optimal training of fair predictive models. In Conference on Causal Learning and Reasoning (pp. 594-617). PMLR.<\/span><\/section>\n

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

Causal determinants of postoperative length of stay in cardiac surgery using causal graphical learning<\/span><\/strong><\/span><\/h4>\n
Daniel Malinsky, Columbia University<\/span><\/section>\n
\u56e0\u679c\u63a8\u7406\u4e2d\u7684\u8bb8\u591a\u76ee\u6807\uff0c\u5305\u62ec\u4f30\u8ba1\u5e73\u5747\u6cbb\u7597\u6548\u679c\u548c\u7406\u89e3\u7279\u5b9a\u8def\u5f84\u673a\u5236\uff0c\u53d6\u51b3\u4e8e\u4e86\u89e3\u67d0\u4e2a\u9886\u57df\u7684\u5b9a\u6027\u56e0\u679c\u7ed3\u6784\u3002\u5728\u8fd9\u9879\u5de5\u4f5c\u4e2d\uff0c\u6211\u4eec\u5c06\u56fe\u5f62\u56e0\u679c\u53d1\u73b0\u65b9\u6cd5\uff08\u7279\u522b\u662fFCI\u7b97\u6cd5\uff09\u5e94\u7528\u4e8e\u7ea6\u7ff0\u00b7\u970d\u666e\u91d1\u65af\u533b\u9662\u7535\u5b50\u5065\u5eb7\u8bb0\u5f55\uff08EHR\uff09\u5f62\u5f0f\u7684\u89c2\u5bdf\u6570\u636e\u3002\u6211\u4eec\u7684\u76ee\u6807\u662f\u4e86\u89e3\u5fc3\u810f\u624b\u672f\u60a3\u8005\u672f\u540e\u4f4f\u9662\u65f6\u95f4\u7684\u56e0\u679c\u51b3\u5b9a\u56e0\u7d20\uff0c\u4ee5\u4fbf\u4e3a\u652f\u6301\u60a3\u8005\u66f4\u5feb\u5eb7\u590d\u7684\u53ef\u80fd\u5e72\u9884\u63aa\u65bd\u63d0\u4f9b\u4fe1\u606f\u3002\u5bf9\u4e8e2011\u5e74\u81f32016\u5e74\u671f\u95f4\u63a5\u53d7\u9694\u79bbCABG\u6216\u9694\u79bbAVR\u624b\u672f\u7684\u60a3\u8005\uff0c\u6211\u4eec\u4f7f\u7528\u4e86\u7535\u5b50\u5065\u5eb7\u8bb0\u5f55\u6570\u636e\u7684\u56e0\u679c\u56fe\u5f62\u5efa\u6a21\u3002\u5728\u7ed9\u5b9a\u4e34\u5e8a\u80cc\u666f\u77e5\u8bc6\u7684\u60c5\u51b5\u4e0b\uff0c\u4f7f\u7528\u6765\u81eaTetrad\u8f6f\u4ef6\u7684\u5feb\u901f\u56e0\u679c\u63a8\u7406\uff08FCI\uff09\u7b97\u6cd5\u5bf9\u6570\u636e\u8fdb\u884c\u4f30\u8ba1\uff0c\u4ee5\u63cf\u8ff0pLOS\u7684\u76f4\u63a5\u548c\u95f4\u63a5\u539f\u56e0\u3002\u7136\u540e\uff0c\u6211\u4eec\u4f7f\u7528\u6f5c\u5728\u53d8\u91cfIDA\u7b97\u6cd5\u6765\u4f30\u8ba1\u611f\u5174\u8da3\u56e0\u679c\u6548\u5e94\u7684\u5f3a\u5ea6\u3002\u6700\u540e\uff0c\u6211\u4eec\u5bf9pLOS\u8fdb\u884c\u4e86\u4e00\u6b21\u7ebf\u6027\u56de\u5f52\uff0c\u4ee5\u5c06\u7edf\u8ba1\u5173\u8054\u4e0e\u56e0\u679c\u5206\u6790\u6240\u83b7\u5f97\u7684\u7ed3\u679c\u8fdb\u884c\u5bf9\u6bd4\u3002\u57fa\u4e8e\u6211\u4eec\u7684\u65b9\u6cd5\uff0c\u6211\u4eec\u53d1\u73b0\u4f7f\u7528\u56de\u987e\u6027EHR\u6570\u636e\u548c\u80cc\u666f\u4e34\u5e8a\u77e5\u8bc6\uff0c\u56e0\u679c\u56fe\u5f62\u5efa\u6a21\u68c0\u7d22\u4e86pLOS\u7684\u76f4\u63a5\u548c\u95f4\u63a5\u539f\u56e0\u53ca\u5176\u76f8\u5bf9\u5f3a\u5ea6\u3002\u8fd9\u4e9b\u89c1\u89e3\u5c06\u6709\u52a9\u4e8e\u8bbe\u8ba1\u4e34\u5e8a\u65b9\u6848\u548c\u6539\u5584\u60a3\u8005\u7ba1\u7406\u3002\u6b64\u5916\uff0c\u6211\u4eec\u8fd8\u8ba8\u8bba\u4e86\u5c06\u56e0\u679c\u53d1\u73b0\u65b9\u6cd5\u5e94\u7528\u4e8e\u7535\u5b50\u5065\u5eb7\u8bb0\u5f55\u7684\u6311\u6218\u548c\u672a\u6765\u5de5\u4f5c\u7684\u673a\u4f1a\u3002<\/span><\/section>\n
\u53c2\u8003\u6587\u732e\uff1a<\/span><\/section>\n
Lee, J. J., Srinivasan, R., Ong, C. S., Alejo, D., Schena, S., Shpitser, I., … & Malinsky, D. (2022). Causal Determinants of Postoperative Length of Stay in Cardiac Surgery using Causal Graphical Learning. The Journal of Thoracic and Cardiovascular Surgery.<\/span><\/section>\n

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


<\/h2>\n

\n
\n
\n
\n
\n

<\/section>\n

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

\u4e13\u9898\u5341\u4e00\uff08Peng Cui\uff0cChair\uff09<\/span><\/strong><\/span><\/span><\/strong><\/strong><\/p>\n<\/section>\n

\n

<\/section>\n

<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/h3>\n


<\/p>\n

Learning Causality with Graphs<\/span><\/strong><\/span><\/h4>\n
Jundong Li  University of Virginia<\/span><\/section>\n
\u5b66\u4e60\u56e0\u679c\u6027\u662f\u4eba\u5de5\u667a\u80fd\u4e00\u9879\u91cd\u8981\u7684\u80fd\u529b\uff0c\u5e76\u4e14\u53ef\u4ee5\u4f5c\u4e3a\u4eba\u5de5\u667a\u80fd\u7684\u57fa\u7840\u3002\u5728\u56e0\u679c\u5b66\u4e60\u4e2d\uff0c\u4e00\u9879\u57fa\u672c\u7684\u95ee\u9898\u662f\uff0c\u5982\u4f55\u7406\u89e3\u4e00\u4e2a\u7279\u5b9a\u7684\u5e72\u9884\u5bf9\u4e00\u4e2a\u91cd\u8981\u7684\u7ed3\u679c\u7684\u4f5c\u7528\uff0c\u7279\u522b\u662f\u5728\u5404\u79cd\u91cd\u8981\u7684\u9886\u57df\uff0c\u4f8b\u5982\u5065\u5eb7\u76d1\u6d4b\u3001\u6559\u80b2\u3001\u7535\u5b50\u5546\u52a1\u7b49\u7b49\u3002\u89e3\u51b3\u8fd9\u4e00\u95ee\u9898\u7684\u4e00\u79cd\u6d41\u884c\u65b9\u6cd5\u662f\u76f4\u63a5\u4f7f\u7528\u89c2\u5bdf\u6027\u7684\u6570\u636e\uff0c\u56e0\u4e3a\u968f\u673a\u8bd5\u9a8c\u5f80\u5f80\u975e\u5e38\u6602\u8d35\uff0c\u8d39\u65f6\uff0c\u5e76\u4e14\u53ef\u80fd\u5728\u4e00\u4e9b\u60c5\u51b5\u4e0b\u9762\u4e34\u4f26\u7406\u95ee\u9898\u3002\u7136\u800c\uff0c\u73b0\u5b58\u7684\u6570\u636e\u9a71\u52a8\u7684\u65b9\u6cd5\u5f80\u5f80\u6709\u4ee5\u4e0b\u5c40\u9650\u6027\uff1a\uff081\uff09\u5047\u8bbe\u89c2\u5bdf\u6027\u6570\u636e\u662f\u72ec\u7acb\u540c\u5206\u5e03\u7684\uff0c\u751a\u81f3\u5047\u8bbe\u5404\u4e2a\u5355\u5143\u4e4b\u95f4\u662f\u6ca1\u6709\u76f8\u4e92\u5f71\u54cd\u7684\uff1b\uff082\uff09\u5ffd\u7565\u4e86\u6f5c\u5728\u7684\u6df7\u6742\u7684\u5f71\u54cd\uff1b\u540c\u65f6\uff0c\u771f\u5b9e\u4e16\u754c\u6570\u636e\u5e38\u5e38\u662f\u76f8\u4e92\u8fde\u63a5\u7684\uff0c\u5e76\u4e14\u53ef\u4ee5\u88ab\u62bd\u8c61\u4e3a\u56fe\uff0c\u4f8b\u5982\u793e\u4ea4\u7f51\u7edc\uff0c\u751f\u7269\u7f51\u7edc\uff0c\u77e5\u8bc6\u56fe\u8c31\u7b49\u3002\u65e0\u5904\u4e0d\u5728\u7684\u56fe\u7684\u6570\u636e\u4e5f\u7ed9\u6211\u4eec\u5e26\u6765\u4e86\u63a7\u5236\u6f5c\u5728\u6df7\u6742\u5f71\u54cd\u3001\u6784\u5efa\u66f4\u6709\u6548\u7684\u6a21\u578b\u7684\u673a\u4f1a\uff0c\u4ece\u800c\u5e26\u6765\u65e0\u504f\u7684\u56e0\u679c\u4f5c\u7528\u4f30\u8ba1\u3002\u5728\u8fd9\u4e00\u8bb2\u5ea7\u4e2d\uff0cJundong Li\u8001\u5e08\u4ecb\u7ecd\u4e86\u73b0\u9636\u6bb5\u4f7f\u7528\u56fe\u6a21\u578b\u5b66\u4e60\u56e0\u679c\u4f5c\u7528\u7684\u6709\u5173\u5de5\u4f5c\u3002\u7279\u522b\u5730\uff0cJundong Li\u8bd5\u56fe\u56de\u7b54\u4ee5\u4e0b\u7684\u79d1\u7814\u95ee\u9898\uff1a\u5982\u4f55\u7528\u89c2\u5bdf\u6027\u7684\u6570\u636e\u6765\u83b7\u53d6\u56fe\u7684\u4fe1\u606f\u4ece\u800c\u5b66\u4e60\u56e0\u679c\u4f5c\u7528\uff1f\u5728\u56fe\u4e0d\u65ad\u66f4\u65b0\u7684\u60c5\u51b5\u4e0b\uff0c\u5982\u4f55\u5229\u7528\u5386\u53f2\u4fe1\u606f\u6765\u63a7\u5236\u6f5c\u5728\u6df7\u6742\u7684\u5f71\u54cd\uff0c\u4ece\u800c\u66f4\u597d\u5730\u5b66\u4e60\u56e0\u679c\u4f5c\u7528\uff1f<\/span><\/section>\n

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

Emulating Clinical Trials with Large Scale Electronic Health Records<\/span><\/strong><\/span><\/h4>\n
Fei Wang Cornell University<\/span><\/section>\n
\u836f\u7269\u7684\u53d1\u73b0\u548c\u5f00\u53d1\u662f\u8017\u65f6\u8017\u529b\u7684\u8fc7\u7a0b\uff0c\u63d0\u5347\u4e34\u5e8a\u8bd5\u9a8c\u7684\u6548\u7387\u548c\u51c6\u786e\u6027\u662f\u5236\u836f\u4ea7\u4e1a\u7684\u91cd\u4e2d\u4e4b\u91cd\u3002\u8fd1\u4e9b\u5e74\u6765\uff0c\u5927\u4f30\u6478\u7684\u771f\u5b9e\u4e16\u754c\u75c5\u4f8b\u6570\u636e\uff0c\u7279\u522b\u662f\u7eb5\u5411\u7684\u7535\u5b50\u75c5\u5386\u8bb0\u5f55\u53d7\u5230\u4e86\u5173\u6ce8\u3002\u8fd9\u4e9b\u6570\u636e\u5305\u542b\u4e86\u75c5\u4f8b\u7684\u6cbb\u7597\u7684\u6709\u6548\u6027\u548c\u5b89\u5168\u6027\u7684\u6709\u7528\u4fe1\u606f\u3002\u5bf9\u8fd9\u4e9b\u4fe1\u606f\u7684\u6709\u6548\u6316\u6398\u53ef\u4ee5\u5e2e\u52a9\u6211\u4eec\u66f4\u597d\u5730\u6307\u5bfc\u4e34\u5e8a\u8bd5\u9a8c\u7684\u8bbe\u8ba1\u3002\u4e34\u5e8a\u8bd5\u9a8c\u4eff\u771f\u6307\u7684\u662f\u7528\u771f\u5b9e\u4e16\u754c\u75c5\u4f8b\u6570\u636e\u6765\u6a21\u4eff\u4e34\u5e8a\u8bd5\u9a8c\u7684\u8fc7\u7a0b\u3002\u7531\u4e8e\u8fd9\u771f\u5b9e\u4e16\u754c\u6570\u636e\u56de\u987e\u6027\u7684\u7279\u5f81\uff0c\u6211\u4eec\u9700\u8981\u4ed4\u7ec6\u5730\u63a7\u5236\u53ef\u80fd\u5b58\u5728\u7684\u6df7\u6742\u56e0\u7d20\uff0c\u4ece\u800c\u4f30\u8ba1\u4e2a\u4f53\u56e0\u679c\u4f5c\u7528\u3002\u5728\u8fd9\u4e00\u8bb2\u5ea7\u4e2d\uff0cFei Wang\u8001\u5e08\u8bb2\u8ff0\u4e86\u4ed6\u8fd1\u671f\u5229\u7528\u6a21\u62df\u5b9e\u9a8c\u6765\u8bc6\u522b\u6cbb\u7597\u963f\u5c14\u5179\u6d77\u9ed8\u75c7\u6709\u6548\u836f\u7269\u7684\u7814\u7a76\uff0c\u4ee5\u53ca\u5229\u7528\u6a21\u62df\u5b9e\u9a8c\u6765\u8bc6\u522b\u65b0\u51a0\u80ba\u708e\u611f\u67d3\u540e\u9057\u75c7\u7684\u6f5c\u5728\u75c7\u72b6\u7684\u7814\u7a76\uff0c\u5e76\u4e14\u6307\u51fa\u4e86\u76f8\u5e94\u7684\u6311\u6218\u548c\u672a\u6765\u7684\u7814\u7a76\u65b9\u5411\u3002<\/span><\/section>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

\u5bfc\u8bed \u4e3a\u4e86\u66f4\u591a\u5730\u63a8\u52a8\u56e0\u679c\u63a8\u65ad\u5b66\u79d1\u7684\u53d1\u5c55\uff0c\u805a\u96c6\u56fd\u5185\u5916\u56e0\u679c\u63a8\u65ad\u7684\u4e00\u7ebf\u79d1\u7814\u5de5\u4f5c\u8005\uff0c\u5171\u540c\u8ba8\u8bba\u56e0\u679c\u79d1\u5b66\u7684\u6700\u65b0\u8fdb\u5c55\uff0c\u5317\u4eac\u5927\u5b66\u8bb2\u5e2d\u6559\u6388\u3001\u5317\u4eac\u5927\u5b66\u516c\u5171\u536b\u751f\u5b66\u9662\u751f\u7269\u7edf\u8ba1\u7cfb\u4e3b\u4efb\u3001\u5317\u4eac\u5927\u5b66\u5317\u4eac\u56fd\u9645\u6570\u5b66\u7814\u7a76\u4e2d\u5fc3\u751f\u7269\u7edf\u8ba1\u548c\u4fe1\u606f\u7814\u7a76\u5ba4\u4e3b\u4efb\u5468\u6653\u534e\u7b49\u53d1\u8d77\u4e86\u6cdb\u592a\u5e73\u6d0b\u56e0\u679c\u63a8\u65ad\u5927\u4f1a\uff0c\u4f1a\u8bae\u4e8e2021\u5e749\u670817\u65e5\u81f318\u65e5\u4e3e\u884c\u3002\u6cdb\u592a\u5e73\u6d0b\u56e0\u679c\u63a8…<\/p>\n","protected":false},"author":1,"featured_media":37591,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[],"special":[],"_links":{"self":[{"href":"https:\/\/swarma.org\/index.php?rest_route=\/wp\/v2\/posts\/37602"}],"collection":[{"href":"https:\/\/swarma.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/swarma.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/swarma.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/swarma.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=37602"}],"version-history":[{"count":0,"href":"https:\/\/swarma.org\/index.php?rest_route=\/wp\/v2\/posts\/37602\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/swarma.org\/index.php?rest_route=\/wp\/v2\/media\/37591"}],"wp:attachment":[{"href":"https:\/\/swarma.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=37602"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/swarma.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=37602"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/swarma.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=37602"},{"taxonomy":"special","embeddable":true,"href":"https:\/\/swarma.org\/index.php?rest_route=%2Fwp%2Fv2%2Fspecial&post=37602"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}