← 返回起始笔记位置

知识卡片

结构图卡 · 8967.f · 生成模型:GPT-6

树上边缘推断的数据路线

理解等级:中级

结构图:

flowchart TD
 A[定义局部证据与两两势] --> B[固定观测并检查状态空间]
 B --> C[叶向根汇集求和消息]
 C --> D[根向叶传播反向消息]
 D --> E[局部势乘全部入站消息]
 E --> F[按节点归一化边缘]
 F --> G[按损失选择点估计]

该路线的精确性针对给定树形模型,不能自动延伸为一般有环图的保证。

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

      For more complicated graphical models, we want to replace the manual factorization above with an automatic procedure for identifying the computations needed for marginalizations and to cache them efficiently. BP does that by identifying those reusable sums, that is, the messages.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第29章《概率图模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c029.xhtml) - 结论依据:摘录中关于树上边缘推断的数据路线的定义、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:For more complicated graphical models, we want to replace the manual factorization above with an automatic procedure for identifying the computations needed for marginalizations and to cache them efficiently. BP does that by identifying those reusable sums, that is, the messages.