← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8994 · 生成模型:GPT-6

条件后验与聚合分布不同

理解等级:高级

qψ(z|x)描述单个观测的潜不确定性,qψ(z)=E数据qψ(z|x)是汇总全部观测后的分布。对每个条件后验施加KL会间接影响聚合分布,但二者不是同一个约束。编码均值的分布也不同于从q抽样的分布。

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

      The KL term encourages the encodings z(i) to be near a unit Gaussian distribution. To get an intuition for the effect of this term, consider the case where we fix to be 1; this is still a valid model for qψ(Z|x), just with lower capacity because it has fewer free parameters. In this simple case, the KL term reduces to . The effect of this

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第33章《生成建模与表示学习》(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/c033.xhtml) - 结论依据:摘录中关于条件后验与聚合分布不同的定义、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:The KL term encourages the encodings z(i) to be near a unit Gaussian distribution. To get an intuition for the effect of this term, consider the case where we fix to be 1; this is still a valid model for qψ(Z|x), just with lower capacity because it has fewer free parameters. In this simple case, the KL term reduces to . The effect of this