A structural characterization of shortcut features for prediction.
A structural characterization of shortcut features for prediction.
复制标题
快捷特征的结构表征用于预测。
DOI：
10.1007/s10654-022-00892-3
复制
发表时间：
2022-06
影响因子：
13.6
通讯作者：
Beam, Andrew
中科院分区：
文献类型：
--
作者：
Bellamy, David;Hernan, Miguel A.;Beam, Andrew
With the rising use of machine learning for healthcare applications, practitioners are increasingly confronted with the limitations of prediction models that are trained in one setting but meant to be deployed in several others. One recently identified limitation is so-called shortcut learning, whereby a model learns to associate features with the prediction target that do not maintain their relationship across settings. Famously, the watermark on chest x-rays has been demonstrated to be an instance of a shortcut feature. In this viewpoint, we attempt to give a structural characterization of shortcut features in terms of causal DAGs. This is the first attempt at defining shortcut features in terms of their causal relationship with a model’s prediction target.
登录
查看更多内容
影响因子：
64.8
作者：
Esteva A;Kuprel B;Novoa RA;Ko J;Swetter SM;Blau HM;Thrun S
通讯作者：
Thrun S
影响因子：
120.7
作者：
Gulshan, Varun;Peng, Lily;Webster, R.
通讯作者：
Webster, R.
影响因子：
1.4
作者：
Pearl, Judea
通讯作者：
Pearl, Judea
DOI：
10.1016/j.spinee.2020.08.012
发表时间：
2021-10
期刊：
The spine journal : official journal of the North American Spine Society
影响因子：
--
作者：
Schmaltz A;Beam AL
通讯作者：
Beam AL
影响因子：
15.2
作者：
Badgeley, Marcus A.;Zech, John R.;Dudley, Joel T.
通讯作者：
Dudley, Joel T.