← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 9173.b · 生成模型:GPT-6

卷积示例先核对翻核约定

理解等级:中级

书中10.23对窗口与同索引核直接点乘,属于常见神经网络实现的相关形式,10.22却写数学卷积。对可学习核可通过约定适配,但比较手算导数核时符号不同;实现应明确操作而非仅靠“卷积”名称。

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

      This operation is therefore simply computing the dot product of the value of nine pixels with the kernel weights. Adding a bias value and an activation function such as ReLu is therefore identical to adding a hidden layer with nine neurons!

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第10章《人工神经网络》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_10.xhtml) - 结论依据:摘录中关于卷积示例先核对翻核约定的定义、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:This operation is therefore simply computing the dot product of the value of nine pixels with the kernel weights. Adding a bias value and an activation function such as ReLu is therefore identical to adding a hidden layer with nine neurons!