知识卡片
采样与寻找众数不同
从密度或能量模型生成样本,需要按其分布抽样;优化寻找最高密度或最低能量点,只得到众数附近解。反复返回同一个高分点会丢失分布多样性。间接生成路线应明确使用什么采样算法,而非把优化最高分直接当作正确采样。
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
- · 生成模型
原书出处 · 1 处
- · 查看原书依据
2. Indirect approach: learn a function E : X → ℝ and generate samples by finding values for x that score highly under this function. In the generative modeling literature, the direct approach is sometimes called an implicit model. This is because the probability density is never explicitly represented. However, note that this “implicit” model explicitly describes the way data is generated. To avoid confusion,
参考来源
- 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第32章《生成模型》(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/c032.xhtml)
- 结论依据:摘录中关于采样与寻找众数不同的定义、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:2. Indirect approach: learn a function E : X → ℝ and generate samples by finding values for x that score highly under this function. In the generative modeling literature, the direct approach is sometimes called an implicit model. This is because the probability density is never explicitly represented. However, note that this “implicit” model explicitly describes the way data is generated. To avoid confusion,