知识卡片

普通读书笔记卡 · 6713.d · 生成模型:GPT-6

核翻转决定卷积与互相关区别

理解等级:中级

数学卷积会按定义翻转一个序列,通常具有交换性;CNN实现常用互相关,不必具备相同对称关系。对称核上二者结果可能一致,因此仅用[0.5,0.5]或对称低高低例不能说明一般情形没有区别。

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

      In figure 7.6, you can think of the cartoon fence as a kernel of alternating numerical values. The zebra in the background is like your data, with alternating numerical values for the light and dark areas in its stripes. And because multiplication and addition are commutative operations, convolution is symmetric, so if you prefer, you can think of the zebra stripes as the filter and

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第07章《卷积网络识别文本模式》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-7.html) - 结论依据:摘录中关于核翻转决定卷积与互相关区别的定义、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:In figure 7.6, you can think of the cartoon fence as a kernel of alternating numerical values. The zebra in the background is like your data, with alternating numerical values for the light and dark areas in its stripes. And because multiplication and addition are commutative operations, convolution is symmetric, so if you prefer, you can think of the zebra stripes as the filter and