知识卡片
卷积层结合局部连接与权重共享
一个输出位置读取局部窗口并用共享核加权,扫过所有位置形成特征图,多个输出通道学习不同核。3×3单通道窗口对应一个有9个输入的输出单元,原文将其称为隐藏层9个神经元不准确;多输入通道时核还跨通道求和。
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
- · 人工神经网络
原书出处 · 1 处
- · 查看原书依据
Performing the convolution by moving the convolution kernel with a width of (2r + 1) across an entire X ×Y image is therefore akin to creating (X − 2r)(Y− 2r) “convolutional” neurons; the resulting structure is called a feature map. Note that the “weights” of the feature map—that is, the entries of the kernel matrix—are identical for each neuron in the feature map. We can
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:摘录中关于卷积层结合局部连接与权重共享的定义、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:Performing the convolution by moving the convolution kernel with a width of (2r + 1) across an entire X ×Y image is therefore akin to creating (X − 2r)(Y− 2r) “convolutional” neurons; the resulting structure is called a feature map. Note that the “weights” of the feature map—that is, the entries of the kernel matrix—are identical for each neuron in the feature map. We can