知识卡片

普通读书笔记卡 · 6538.e · 生成模型:GPT-6

变分推断优化近似分布

理解等级:中级

变分推断选择一个带参数的近似分布,用优化让它接近目标后验。近似族、目标函数和优化过程会影响误差,自动选择guide可减少操作负担,却不保证覆盖真实后验的所有形态或找到全局最佳近似。

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

      approximating distribution” that resembles the target distribution. That stochastic process has parameters that we optimize using gradient-based techniques now common in deep learning software. The objective function of the optimization tries to minimize the difference between the approximating distribution and the target distribution.

    参考来源

    - 位置:《Causal AI》第02章《概率生成建模入门》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-2.xhtml) - 结论依据:摘录中关于变分推断优化近似分布的定义、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:approximating distribution” that resembles the target distribution. That stochastic process has parameters that we optimize using gradient-based techniques now common in deep learning software. The objective function of the optimization tries to minimize the difference between the approximating distribution and the target distribution.