知识卡片
计算图与概率图含义不同
神经网络图通常描述确定的张量计算依赖,概率图描述随机变量的分布分解与条件独立,不能仅因画法相似就直接对应边的语义。二者可以结合:神经网络可参数化条件分布或势函数,也可学习消息更新;是否输出概率取决于模型和校准。
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
- · 概率图模型
原书出处 · 1 处
- · 查看原书依据
Nodes: The nodes of a NN diagram typically represent a real-valued scalar activations. The nodes of a PGM represent a random variable, real or discrete-valued, and scalars or vectors. Edges: Edges in PGMs determine statistical dependencies. Edges in NNs determine which are involved in a linear summation to the next layer. Computation: The big difference between PGMs and NNs lies in what they calculate. PGMs
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:摘录中关于计算图与概率图含义不同的定义、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:Nodes: The nodes of a NN diagram typically represent a real-valued scalar activations. The nodes of a PGM represent a random variable, real or discrete-valued, and scalars or vectors. Edges: Edges in PGMs determine statistical dependencies. Edges in NNs determine which are involved in a linear summation to the next layer. Computation: The big difference between PGMs and NNs lies in what they calculate. PGMs