知识卡片

结构图卡 · 6574.f · 生成模型:GPT-6

双世界结构把对象联系起来

理解等级:中级

结构图:

flowchart TD
 NX[共享外生NX] --> X[事实X]
 NY[共享外生NY] --> Y[事实Y]
 X --> Y
 NY --> Yc[替代结果Y_x]
 C[假想世界X固定为x] --> Yc

图示无隐藏混杂的两变量SCM。事实Y和替代Y_x共用NY,假想X的原入边被切断;这条跨世界联系来自所选结构假设,不是由两个边缘分布单独恢复。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      We’ll suppose that the original nodes of the DAG are the endogenous variables of the SCM and expand the DAG visualization by adding the exogenous variables. Further, the two causal DAGs will use the same exogenous nodes. We call the resulting graph a parallel world graph (or, for this typical case of two possible worlds, a “twin-world graph”). Figure 8.9 visualizes the parallel world graph.

    参考来源

    - 位置:《Causal AI》第08章《反事实与平行世界》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-8.xhtml) - 结论依据:摘录中关于双世界结构把对象联系起来的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:We’ll suppose that the original nodes of the DAG are the endogenous variables of the SCM and expand the DAG visualization by adding the exogenous variables. Further, the two causal DAGs will use the same exogenous nodes. We call the resulting graph a parallel world graph (or, for this typical case of two possible worlds, a “twin-world graph”). Figure 8.9 visualizes the parallel world graph.