知识卡片
相关代理可能随机制改变失效
若A→B→C,利用C预测A可以在既定机制下有效;当A到B的机制改变时,原相关模式可能不再保持。分析应定位预测所依赖的机制及其变化条件,而不是把高预测准确率直接当作稳健因果理解。显式建模也必须有可信假设才能支持迁移。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
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
assume a chain of events: A causes B causes C. If a model, trained on this data, has learned to predict A from C, it is inherently fragile. It will fail completely if, in a new setting, the link between A and B is somehow broken. Again, explicit causal models tackle this issue directly, greatly increasing the chances that a performant model will generalize well.
参考来源
- 位置:《Causal AI》第00章《序言、前言与阅读指南》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/foreword.xhtml)
- 结论依据:摘录中关于相关代理可能随机制改变失效的定义、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:assume a chain of events: A causes B causes C. If a model, trained on this data, has learned to predict A from C, it is inherently fragile. It will fail completely if, in a new setting, the link between A and B is somehow broken. Again, explicit causal models tackle this issue directly, greatly increasing the chances that a performant model will generalize well.