知识卡片

普通读书笔记卡 · 6575.d · 生成模型:GPT-6

图中确定性破坏的是忠实反推

理解等级:高级

含复制确定变量的平行世界模型仍可满足DAG马尔可夫性质,d分离的独立性蕴含并不因确定性一般失效。真正要避免的是把d连接必然解释为依赖,或用图完整捕获所有由确定关系带来的独立。原文把这种限制概括为d分离不可靠,范围过宽。

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

      d-connection to reason about the dependence between endogenous variables across worlds. That’s because the law of consistency requires that the same endogenous variables across worlds must have the same value (unless one of the pairs is impacted by an intervention). Two variables always having the same value is a perfect dependence; the rules of d-separation do not capture that dependence on the parallel world graph.

    参考来源

    - 位置:《Causal AI》第09章《通用反事实推断算法》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-9.xhtml) - 结论依据:摘录中关于图中确定性破坏的是忠实反推的定义、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:d-connection to reason about the dependence between endogenous variables across worlds. That’s because the law of consistency requires that the same endogenous variables across worlds must have the same value (unless one of the pairs is impacted by an intervention). Two variables always having the same value is a perfect dependence; the rules of d-separation do not capture that dependence on the parallel world graph.